<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Justin Bowen on Medium]]></title>
        <description><![CDATA[Stories by Justin Bowen on Medium]]></description>
        <link>https://medium.com/@tonsoffun111?source=rss-5e886beaba05------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*0xPGeD21svhUPNMP0kvT8g@2x.jpeg</url>
            <title>Stories by Justin Bowen on Medium</title>
            <link>https://medium.com/@tonsoffun111?source=rss-5e886beaba05------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 08:23:13 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@tonsoffun111/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Autonomous Plant Feeding Open Source Prototype]]></title>
            <link>https://medium.com/greenthumbio/autonomous-plant-feeding-open-source-prototype-df73972e570e?source=rss-5e886beaba05------2</link>
            <guid isPermaLink="false">https://medium.com/p/df73972e570e</guid>
            <category><![CDATA[raspberry-pi]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[arduino]]></category>
            <category><![CDATA[gardening]]></category>
            <category><![CDATA[automation]]></category>
            <dc:creator><![CDATA[Justin Bowen]]></dc:creator>
            <pubDate>Wed, 06 Mar 2019 23:02:32 GMT</pubDate>
            <atom:updated>2019-03-09T00:57:59.654Z</atom:updated>
            <content:encoded><![CDATA[<h3>Autonomous Plant Feeding Open-Source Prototype</h3><h4>How to build 3rd party software adapters for hardware you can’t afford</h4><h3>Developing Adapters For 3rd Party Products</h3><p>GreenThumb is working to build a platform that connects growers to the systems they’re already comfortable using today. This blog post will read a bit like a tutorial on how to DIY your own dosing system.</p><h4>Peristaltic Dosing Pump Solution</h4><p>I’m not trying to compete with commercial dosing system here, but I will be demonstrating a cost effective solution to control the mixing of solutions autonomously. In doing this we also define the features and components required for a base adapter than can be augmented to connect with any 3rd party offering. Enough of the diatribe, let’s jump in, first to assess the industry standard solution.</p><h3>Theirs $3300: For Commercial Production Deployments</h3><h4>Connect pH Controller $500</h4><p>This device is a commercial grade IoT pH control unit with liquid temperature and pH probes and a peristaltic pump for dosing pH buffer solution at a rate of 10 ml per minute.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SNKRYW2b0QgrXgTYQAc6FQ.jpeg" /><figcaption>Controller ✅ peristaltic pump ✅ pH probe ✅ temperature probe ✅ find it on <a href="https://www.amazon.com/gp/product/B00QV852OO/">Amazon Prime for $427</a></figcaption></figure><h4>Microdoser $800</h4><p><a href="https://www.amazon.com/gp/product/B0735HB8RT/">Bluelab PeriPod M3</a></p><h4>Macrodoser $2000</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/500/1*sdjPPcaJjevADT1c73gIFQ.jpeg" /><figcaption>With 3 peristaltic pumps calibrated to move 1200ml of solution per min this well packaged device can be yours for a cool <a href="https://www.amazon.com/gp/product/B0735FTV98/">$1900 on Amazon</a> 🙀 FREE Shipping though! 📦</figcaption></figure><h3>Ours $400: For Testing and Development Purposes</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dO_ELyiOZ4avqoPy44D_-g.jpeg" /></figure><h4>Materials</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*6ClYaaJSe3zT6OW0FwzA2g.jpeg" /><figcaption>This might be gross, but it works!</figcaption></figure><p>$160 <a href="https://www.amazon.com/Atlas-Scientific-pH-Kit-0-14/dp/B00798612C/ref=sr_1_3?ie=UTF8&amp;qid=1550352936&amp;sr=8-3&amp;keywords=atlas+scientific+ph+kit">pH Sensor Kit</a></p><p>$40 <a href="https://www.amazon.com/ELEMENT-Element14-Raspberry-Pi-Motherboard/dp/B07BDR5PDW/ref=sr_1_3?s=pc&amp;ie=UTF8&amp;qid=1550353014&amp;sr=1-3&amp;keywords=raspberry+pi+3+b%2B">Raspberry Pi 3 B+</a></p><p>$8 <a href="https://www.amazon.com/Samsung-MicroSD-Adapter-MB-ME32GA-AM/dp/B06XWN9Q99/ref=pd_bxgy_147_img_2/147-0723387-8478817?_encoding=UTF8&amp;pd_rd_i=B06XWN9Q99&amp;pd_rd_r=ff6c9313-3232-11e9-aa96-e74d7a705039&amp;pd_rd_w=juJhW&amp;pd_rd_wg=y6NzT&amp;pf_rd_p=6725dbd6-9917-451d-beba-16af7874e407&amp;pf_rd_r=ZZH6BX88MKGBGF4P572P&amp;psc=1&amp;refRID=ZZH6BX88MKGBGF4P572P">SD Card</a></p><p>$19 <a href="https://www.amazon.com/gp/product/B071GT7JB4/ref=ppx_yo_dt_b_asin_title_o03__o00_s00?ie=UTF8&amp;psc=1">8-channel USB Relay</a></p><p>$17 <a href="https://www.amazon.com/gp/product/B07GFSGQTG/ref=ppx_yo_dt_b_asin_title_o00__o00_s00?ie=UTF8&amp;psc=1">Power Supply 12V and 5V</a></p><p>$35 <a href="https://www.amazon.com/gp/product/B01DER02OE/ref=ppx_yo_dt_b_asin_title_o04__o00_s01?ie=UTF8&amp;psc=1">2 x Peristaltic Pump 12v 500ml/min</a></p><p>$15 <a href="https://www.amazon.com/Yosoo-Peristaltic-Aquarium-chemical-Analytical/dp/B00VHYO9F0/ref=sr_1_5?keywords=peristaltic+dosing+pump&amp;qid=1551845133&amp;s=gateway&amp;sr=8-5">5 x Peristaltic Pump 6v 60ml/min</a></p><p>$389</p><h4>Optional</h4><p>$9 Soft Tube Fixture for Dosing Pump for 8 Tubes</p><h4>Documentation</h4><p><a href="https://github.com/darrylb123/usbrelay">https://github.com/darrylb123/usbrelay</a></p><h4>Code</h4><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/c10c276bddcc0e09c02d644dda30ccbe/href">https://medium.com/media/c10c276bddcc0e09c02d644dda30ccbe/href</a></iframe><p>You can get you ph meter working with the <a href="https://blog.greenthumb.io/greenthumb-io-sensor-kit-66410c52afee">GreenThumb Sensor Kit</a></p><p><a href="https://medium.com/greenthumbio/greenthumb-io-sensor-kit-66410c52afee">GreenThumb IO Sensor Kit</a></p><h4>Demo</h4><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2Fc2PVNTN-mCc%3Ffeature%3Doembed&amp;url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3Dc2PVNTN-mCc&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2Fc2PVNTN-mCc%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/fe1548b47ad29ee170f599e2ea5c4b1b/href">https://medium.com/media/fe1548b47ad29ee170f599e2ea5c4b1b/href</a></iframe><h3>Conclusion</h3><p>For a commercial cultivator you’re going to want to devise a professional solution using brand products as shown below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NK-KzguayX9GdGtpUb5M8w.png" /><figcaption>Great diagram demonstrating a combination of Etatron micro-doser being used with dosatron injectors and a bluelab ph controller from <a href="https://dilutionsolutions.com/nutrient-delivery-system-doser-pick-dosatron-etatron/">https://dilutionsolutions.com/nutrient-delivery-system-doser-pick-dosatron-etatron/</a></figcaption></figure><p>Once you have your setup or once you’ve build your own GreenThumb will be ready to integrate providing automated controls and notifications with visual and sensory insight into your systems.</p><p>If you have any questions or are interested in the project or using my open-source packages, feel free to leave a comment or ping me on <a href="https://linkedin.com/in/tonsoffun111">LinkedIn</a> or <a href="https://www.twitter.com/tonsoffun111">Twitter</a>, I will be responsive.</p><p><a href="https://greenthumb.io/">GreenThumb</a> is also seeking early clients, partners, and initial funding. Yes, I did the work and paid for this all on my own.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=df73972e570e" width="1" height="1" alt=""><hr><p><a href="https://medium.com/greenthumbio/autonomous-plant-feeding-open-source-prototype-df73972e570e">Autonomous Plant Feeding Open Source Prototype</a> was originally published in <a href="https://medium.com/greenthumbio">GreenThumbIO</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[GreenThumb IO Sensor Kit]]></title>
            <link>https://medium.com/greenthumbio/greenthumb-io-sensor-kit-66410c52afee?source=rss-5e886beaba05------2</link>
            <guid isPermaLink="false">https://medium.com/p/66410c52afee</guid>
            <category><![CDATA[websocket]]></category>
            <category><![CDATA[computer-vision]]></category>
            <category><![CDATA[open-source]]></category>
            <category><![CDATA[iot]]></category>
            <category><![CDATA[arduino]]></category>
            <dc:creator><![CDATA[Justin Bowen]]></dc:creator>
            <pubDate>Wed, 06 Mar 2019 22:41:12 GMT</pubDate>
            <atom:updated>2019-03-06T22:50:48.789Z</atom:updated>
            <content:encoded><![CDATA[<h4>Deploy GreenThumb IO’s Open-Source Sensor Kit</h4><p><a href="https://greenthumb.io/">GreenThumb IO</a> is a cannabis computer vision and AI product. We are open-sourcing our prototypes that we used to develop our platform capable of integrating with any industry standard product and service providers. I have a series of blogs outlining our projects which can be found here on Medium — follow to stay in the loop on new developments!</p><p>When I think of senses I think of the classical five senses: sight, hearing, taste, smell, and touch. These human senses provide inputs for us to navigate situations, allowing us to perform actions based on our perceived environment. Machine sensors provide a wide assortment of opportunities to measure our world.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*LXani_KWZo-Q0o_rLa2HSg.jpeg" /><figcaption>Soil sensors can read temperature and humidity as seen in Adafruit’s Learning post on Wireless Garden Monitoring <a href="https://learn.adafruit.com/wireless-gardening-arduino-cc3000-wifi-modules">https://learn.adafruit.com/wireless-gardening-arduino-cc3000-wifi-modules</a></figcaption></figure><h4>Other Senses</h4><p>Humans also have senses such as balance, which is a culmination of sensory inputs including the <a href="https://en.wikipedia.org/wiki/Vestibular_nerve">inner ear</a>. Accelerometers and gyros are used in <a href="https://www.analog.com/en/applications/markets/automotive-pavilion-home/mems-safety-and-security/electronic-stability-control.html">automotive</a> and <a href="https://www.droneuniversities.com/glossary/electronic-stability-control-esc/">drone</a> Electronic Stability Control systems to provide computers with the inputs required to keep a car hugging a turn and allow a drone to hover.</p><p>Cameras are used in automation to allow cars to <a href="https://towardsdatascience.com/https-medium-com-priya-dwivedi-automatic-lane-detection-for-self-driving-cars-4f8b3dc0fb65">see lanes on highways</a> which can be seen applied in consume products like the Tesla line of Smart Electric Vehicles with <a href="https://www.tesla.com/autopilot">autopilot</a>.</p><h4>Streaming Serial Sensors</h4><p>I’ve put together a backend that includes WebSockets and background jobs to receive and index transmissions from serial into a PostgreSQL database.</p><p><a href="https://github.com/GreenThumbIO/channel-api/">GreenThumbIO/channel-api</a></p><h4>Read Serial Data</h4><p>For every sensor below you can test their data with the python script below.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/b8a692d65f3edb41360a823f9d0bd225/href">https://medium.com/media/b8a692d65f3edb41360a823f9d0bd225/href</a></iframe><p>Now let’s see how we can take readings from each device:</p><h4>Air Temperature &amp; Humidity</h4><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/5faffbd293906692e867882271192040/href">https://medium.com/media/5faffbd293906692e867882271192040/href</a></iframe><h4>Soil Temperature &amp; Humidity</h4><p>This probe can be used to stream soil metrics to enhance your GreenThumb Computer Vision System’s ability to distinguish heat or water stress in the root zone.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/b3be283eb3b99aa620437577be6d0360/href">https://medium.com/media/b3be283eb3b99aa620437577be6d0360/href</a></iframe><h4>Liquid Temperature &amp; pH</h4><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/37066f95fced5e27b2b698ec3567b77c/href">https://medium.com/media/37066f95fced5e27b2b698ec3567b77c/href</a></iframe><h3>References</h3><h4>Computer Vision</h4><ul><li><a href="https://www.learnopencv.com/">Learn OpenCV ( C++ / Python )</a></li><li><a href="https://www.pyimagesearch.com/">PyImageSearch - You can master Computer Vision, Deep Learning, and OpenCV.</a></li><li><a href="https://towardsdatascience.com/https-medium-com-priya-dwivedi-automatic-lane-detection-for-self-driving-cars-4f8b3dc0fb65">Automatic Lane Detection for Self Driving Cars</a></li><li><a href="https://medium.com/greenthumbio/cannabis-leaf-area-analysis-d6449dbda07f">Cannabis Leaf Area Analysis</a></li><li><a href="https://medium.com/greenthumbio/cannabis-bud-area-5ef7433e6f45">Cannabis Bud Area</a></li></ul><h4>Air</h4><p><a href="https://learn.adafruit.com/am2315-encased-i2c-temperature-humidity-sensor">Overview | AM2315 - Encased I2C Temperature/Humidity Sensor | Adafruit Learning System</a></p><h4>Soil</h4><p><a href="https://learn.adafruit.com/wireless-gardening-arduino-cc3000-wifi-modules">Introduction | Wireless Gardening with Arduino + CC3000 WiFi Modules | Adafruit Learning System</a></p><h4>Water Temperature</h4><p><a href="https://learn.adafruit.com/using-ds18b20-temperature-sensor-with-circuitpython">Overview | Using DS18B20 Temperature Sensor with CircuitPython | Adafruit Learning System</a></p><h4>WebSocket Server</h4><p><a href="https://github.com/GreenThumbIO/channel-api/">GreenThumbIO/channel-api</a></p><h4>Computer Audio</h4><p>Voice commands are the new hip user interface</p><ul><li><a href="https://developers.google.com/assistant/sdk/">Google Assistant SDK | Google Assistant SDK | Google Developers</a></li><li><a href="https://developer.amazon.com/alexa-voice-service/sdk">Alexa Voice Service Device SDK</a></li></ul><p>If you have any questions or are interested in the project or using my open-source packages, feel free to leave a comment or ping me on <a href="https://linkedin.com/in/tonsoffun111">LinkedIn</a> or <a href="https://www.twitter.com/tonsoffun111">Twitter</a>, I will be responsive.</p><p><a href="https://greenthumb.io/">GreenThumb</a> is also seeking early clients, partners, and initial funding. Yes, I did the work and paid for this all on my own.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=66410c52afee" width="1" height="1" alt=""><hr><p><a href="https://medium.com/greenthumbio/greenthumb-io-sensor-kit-66410c52afee">GreenThumb IO Sensor Kit</a> was originally published in <a href="https://medium.com/greenthumbio">GreenThumbIO</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Open-Source Smart Cameras]]></title>
            <link>https://medium.com/greenthumbio/open-source-smart-cameras-5b6726918a0e?source=rss-5e886beaba05------2</link>
            <guid isPermaLink="false">https://medium.com/p/5b6726918a0e</guid>
            <category><![CDATA[smart-camera]]></category>
            <category><![CDATA[raspberry-pi]]></category>
            <category><![CDATA[greenthumb-io]]></category>
            <category><![CDATA[cannabis]]></category>
            <category><![CDATA[open-source]]></category>
            <dc:creator><![CDATA[Justin Bowen]]></dc:creator>
            <pubDate>Wed, 06 Mar 2019 22:29:54 GMT</pubDate>
            <atom:updated>2019-03-07T19:42:18.386Z</atom:updated>
            <content:encoded><![CDATA[<h4>Deploy GreenThumb’s Commercial Computer Vision Models</h4><p><a href="https://greenthumb.io/">GreenThumb IO</a> is a cannabis computer vision and AI product. We are open-sourcing our prototypes that we used to develop our platform capable of integrating with any industry standard product and service providers. I have a series of blogs outlining our projects which can be found here on Medium — follow to stay in the loop on new developments!</p><h3>SmartCam.Dev</h3><p>Developing smart cameras is something I’ve been passionate about for years. Now having assisted in the design, development, and deployment of commercial grade smart camera technology in agricultural settings, I have experienced the pains of deploying hundreds of cameras in the field.</p><p>It’s not only a matter of installing the infrastructure to power and network a CCTV system, but to process all the video feeds in real-time, meaning you can’t rely on streaming to the cloud for Computer Vision base AI model inference.</p><p>I will also be launching SmartCam.dev which will host more resources for anyone interested.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*RJRiMFcRC8YbnMtXsjYqFQ.jpeg" /><figcaption><a href="https://commons.wikimedia.org/wiki/File:4X-UHJ_Agridrones_d.jpg">https://commons.wikimedia.org/wiki/File:4X-UHJ_Agridrones_d.jpg</a></figcaption></figure><h4>Edge Computing</h4><p>Drones are actually quite customizable by design from hardware to software that is rugged and robust. All Power Over Ethernet POE CCTV cameras have processors much like your iPhone or Android Smart Phone. Both iOS and Android Operating Systems provide SDKs for developers to deploy their trained Deep Learning models to the App Store using iOS’s <a href="https://developer.apple.com/machine-learning/">Core ML</a> or Android’s <a href="https://developers.google.com/ml-kit/">ML Kit</a>. Smart Cameras like the NestCam iQ have powerful AI features baked in, and do provide an API for developers to retrieve motion captured videos as well as <a href="https://developers.nest.com/reference/api-camera#snapshot_url">periodic snapshots</a>, but there is no SDK for a developer to release their own custom trained model to their App’s user’s cameras. Because they rely on the cloud, NestCams aren’t a viable solution as they cost too much bandwidth.</p><p>Computer Vision-based AI startups are faced with two deployment strategies at the moment: build your own, or deploy servers to process video feeds alongside of legacy dumb POE cameras. This is where what I’ve been working on comes in.</p><p>I’ve made the leap and purchased both SmartCam.dev and SmartCamera.dev in an effort to advance the industry by providing an initial commercially viable smart camera. This site will launch within the coming weeks.</p><p>Below, this is the camera I’ve developed as V1, and it will be available for purchase as well as instructions on how to build your own. If there is significant interest in an Open Smart Camera SDK, I intend to do a crowdfund to make a product that is both visually appealing and industrially viable.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6JFcumArKJiTVXM0D2ZUKQ@2x.jpeg" /><figcaption>Version 1 of our product</figcaption></figure><h3>Hardware &amp; Costs</h3><h4>DIY PoE Deployable Smart Camera —Total Cost: $145</h4><p><a href="https://www.amazon.com/VENTECH-Weatherproof-Aluminum-Surveillance-Enclosure/dp/B015HSSMSQ/">Industrial Weatherproof Aluminum Housing</a> — $27</p><p><a href="https://www.amazon.com/gp/product/B015HSSMSQ/">VENTECH Outdoor Weatherproof Heavy Duty Aluminum CCTV housing Security Surveillance Camera Housing Camera Mount Enclosure with Bracket</a></p><p><a href="https://www.amazon.com/ELEMENT-Element14-Raspberry-Pi-Motherboard/dp/B07BDR5PDW/">Raspberry Pi 3 B+</a> — $40</p><p><a href="https://www.amazon.com/Samsung-MicroSD-Adapter-MB-ME32GA-AM/dp/B06XWN9Q99/">SD Card</a> — $8</p><p><a href="https://www.amazon.com/Raspberry-Pi-NoIR-Camera-Module/dp/B01ER2SMHY">IR Camera Sensor</a> — $30</p><p><a href="https://www.amazon.com/UCTRONICS-802-3af-Splitter-Ethernet-Raspberry/dp/B01MDLUSE7/ref=sr_1_4?keywords=poe+to+usb&amp;qid=1551911167&amp;s=electronics&amp;sr=1-4">PoE to USB</a> — $10</p><p><a href="https://www.amazon.com/UCTRONICS-802-3af-Splitter-Ethernet-Raspberry/dp/B01MDLUSE7/ref=sr_1_4?keywords=poe+to+usb&amp;qid=1551911167&amp;s=electronics&amp;sr=1-4">UCTRONICS IEEE 802.3af Micro USB Active PoE Splitter Power Over Ethernet 48V to 5V 2.4A for Tablets, Dropcam or Raspberry Pi (48V to 5V 2.4A)</a></p><p><a href="https://www.amazon.com/Intel-NCSM2450-DK1-Movidius-Neural-Compute/dp/B076751BN8/ref=sr_1_3?crid=35X6IGCL325C9&amp;keywords=movidius&amp;qid=1551808372&amp;s=electronics&amp;sprefix=movidiu%2Clawngarden%2C194&amp;sr=1-3">Neural Compute Strick VPU</a> (optional) — $75</p><p><a href="https://www.amazon.com/Intel-NCSM2450-DK1-Movidius-Neural-Compute/dp/B076751BN8/ref=sr_1_3?crid=35X6IGCL325C9&amp;keywords=movidius&amp;qid=1551808372&amp;s=electronics&amp;sprefix=movidiu%2Clawngarden%2C194&amp;sr=1-3">Intel NCSM2450.DK1 Movidius Neural Compute Stick</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*a8OhNeaBrF9-JDEh28n3aA@2x.jpeg" /><figcaption>Left: cardboard provides faceplate for Pi Camera Right: PoE adapter powers and provides data to the Pi</figcaption></figure><p>I’ll be providing an STL of a 3D printed faceplate, pi mount, and poe bracket in good time. I’m also working a designing a more compact and contemporary housing which I’ll also provide design files for in the next few months.</p><h3>Software</h3><p>I’m not going to go deep into the software today, but I highly recommend these resources below to get you setup with your Smart Camera project:</p><ul><li><a href="https://github.com/ccrisan/motioneyeos">ccrisan/motioneyeos</a></li><li><a href="https://www.pyimagesearch.com/2018/02/19/real-time-object-detection-on-the-raspberry-pi-with-the-movidius-ncs/">Real-time object detection on the Raspberry Pi with the Movidius NCS - PyImageSearch</a></li><li><a href="https://www.pyimagesearch.com/2015/06/01/home-surveillance-and-motion-detection-with-the-raspberry-pi-python-and-opencv/">Home surveillance and motion detection with the Raspberry Pi, Python, OpenCV, and Dropbox - PyImageSearch</a></li><li><a href="https://medium.com/greenthumbio/cannabis-leaf-area-analysis-d6449dbda07f">Cannabis Leaf Area Analysis</a></li><li><a href="https://medium.com/greenthumbio/cannabis-bud-area-5ef7433e6f45">Cannabis Bud Area</a></li></ul><h3>Conclusion</h3><p>This is just getting started, and I plan to expand on this blog, which is more of a blueprint, into a full featured guide. If you’re interested in contributing to the project please contact me at justin@greenthumb.io.</p><p>This is only V1, V2 will of the camera will look much more sleek. Follow along here on Medium to stay tuned for more in this series of <a href="https://blog.greenthumb.io/greenthumb-opensource-hardware-f88d884f815">Open Source Hardware Project Announcements</a>! I am currently building out a side project, smartcam.dev, which will be the home of an online store and guides to buy or build your own deployable smart cameras.</p><p>If you have any questions or are interested in the project or using my open-source packages, feel free to leave a comment or ping me on <a href="https://linkedin.com/in/tonsoffun111">LinkedIn</a> or <a href="https://www.twitter.com/tonsoffun111">Twitter</a>, I will be responsive.</p><p><a href="https://greenthumb.io/">GreenThumb</a> is also seeking early clients, partners, and initial funding. Yes, I did the work and paid for this all on my own.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5b6726918a0e" width="1" height="1" alt=""><hr><p><a href="https://medium.com/greenthumbio/open-source-smart-cameras-5b6726918a0e">Open-Source Smart Cameras</a> was originally published in <a href="https://medium.com/greenthumbio">GreenThumbIO</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[GreenThumb OpenSource Hardware]]></title>
            <link>https://medium.com/greenthumbio/greenthumb-opensource-hardware-f88d884f815?source=rss-5e886beaba05------2</link>
            <guid isPermaLink="false">https://medium.com/p/f88d884f815</guid>
            <category><![CDATA[greenthumb-io]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[cannabis]]></category>
            <category><![CDATA[computer-vision]]></category>
            <category><![CDATA[open-source]]></category>
            <dc:creator><![CDATA[Justin Bowen]]></dc:creator>
            <pubDate>Wed, 06 Mar 2019 22:05:33 GMT</pubDate>
            <atom:updated>2019-03-09T00:58:30.198Z</atom:updated>
            <content:encoded><![CDATA[<h3>GreenThumb Open-Source Hardware</h3><h4>Project Announcements: Smart Cameras, Sensors, Automated Dosing System For Developers</h4><p><a href="https://greenthumb.io/">GreenThumb IO</a> is a cannabis computer vision and AI product. We are open-sourcing our prototypes that we used to develop our platform capable of integrating with any industry standard product and service providers. I have a series of blogs outlining our projects which can be found here on Medium — follow to stay in the loop on new developments!</p><h3>Smart Cameras</h3><p>Vision is one of our five senses that we use daily to determine when to cross the road or when to water a plant. In <a href="https://medium.com/datadriveninvestor/iot-smart-camera-landscape-942442f88c8">May 2018 I wrote an article</a> detailing the Smart Camera developer kits on the market. The point was to highlight that while these hardware packages are great for desktop tinkering, they’re not ready for commercial deployments in the real world.</p><p><a href="https://medium.com/datadriveninvestor/iot-smart-camera-landscape-942442f88c8">IoT Smart Camera Landscape</a></p><p>Additionally, the traditional CCTV camera manufacturers are either lagging behind or resistant to offering the new technology as a feature offering. It’s now March 2019 and I have yet to see a reasonable offering hit the market.</p><h4>Production Potential</h4><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fplayer.vimeo.com%2Fvideo%2F276867355%3Fapp_id%3D122963&amp;dntp=1&amp;url=https%3A%2F%2Fvimeo.com%2F276867355&amp;image=https%3A%2F%2Fi.vimeocdn.com%2Fvideo%2F709490637_960.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=vimeo" width="960" height="540" frameborder="0" scrolling="no"><a href="https://medium.com/media/1cd6c3e07504fffa3fec49d8cb0643b6/href">https://medium.com/media/1cd6c3e07504fffa3fec49d8cb0643b6/href</a></iframe><p>Though Huawei did announce their concept line of Software Defined CCTV in June of last year, at this time there is<strong> no product </strong>available on the market for Machine Vision models to be deployed in commercial settings.</p><p><a href="https://medium.com/@betabay/how-huawei-is-changing-the-game-with-three-different-technology-innovations-e7dca86ed3e3">How Huawei Is Changing The Game With Three Different Technology Innovations</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/500/1*3RBYQmh9QZtClb7NH1Xm_Q.jpeg" /><figcaption>These are the Dumb Cameras I wish to make smart.</figcaption></figure><p>Because these don’t yet exist I’ve come up with a retro-industrial looking alternative for now:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*a8OhNeaBrF9-JDEh28n3aA@2x.jpeg" /></figure><p><a href="https://www.amazon.com/gp/product/B015HSSMSQ/">VENTECH Outdoor Weatherproof Heavy Duty Aluminum CCTV housing Security Surveillance Camera Housing Camera Mount Enclosure with Bracket</a></p><p>I have a complete blog post that goes over how to build a commercially viable smart camera that you can deploy in any environment.</p><h3>Sensor Kit</h3><p>There are a handful of sensors I’ve found useful in my research. I demonstrate how to stream their readings over WebSockets in a sensor focused blog.</p><h4>GreenThumb Computer Vision System</h4><p><a href="https://medium.com/greenthumbio/greenthumb-io-platform-d6d09ca7fafb">GreenThumb IO Platform</a></p><h4>Air Temperature &amp; Humidity</h4><p>Air probes can be used to stream soil metrics to enhance our <a href="https://greenthumb.io/">GreenThumb</a> Computer Vision System’s ability to distinguish heat or humidity stress in the canopy.</p><p><a href="https://learn.adafruit.com/am2315-encased-i2c-temperature-humidity-sensor">Overview | AM2315 - Encased I2C Temperature/Humidity Sensor | Adafruit Learning System</a></p><h4>Soil Temperature &amp; Humidity</h4><p>Soil probes can be used to stream soil metrics to enhance the <a href="https://greenthumb.io/">GreenThumb</a> Computer Vision System’s ability to distinguish heat or water stress in the root zone.</p><p><a href="https://learn.adafruit.com/wireless-gardening-arduino-cc3000-wifi-modules">Introduction | Wireless Gardening with Arduino + CC3000 WiFi Modules | Adafruit Learning System</a></p><h4>Liquid Temperature &amp; pH</h4><ul><li><a href="https://learn.adafruit.com/using-ds18b20-temperature-sensor-with-circuitpython">Overview | Using DS18B20 Temperature Sensor with CircuitPython | Adafruit Learning System</a></li><li><a href="https://www.atlas-scientific.com/product_pages/kits/ph-kit.html">pH Kit | Atlas Scientific</a></li></ul><h3>Dosing Systems</h3><p>GreenThumb is designed to integrate with industry standard commercial offers that are already out there. Popular brands such as <a href="https://etatron.co.uk/">Etatron</a>, <a href="https://www.bluelab.com/">BlueLab</a>, and <a href="https://www.dosatronusa.com/">Dosatron</a> are prime examples.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NK-KzguayX9GdGtpUb5M8w.png" /><figcaption>Great diagram demonstrating a combination of Etatron micro-doser being used with dosatron injectors and a bluelab ph controller from <a href="https://dilutionsolutions.com/nutrient-delivery-system-doser-pick-dosatron-etatron/">https://dilutionsolutions.com/nutrient-delivery-system-doser-pick-dosatron-etatron/</a></figcaption></figure><p>Preparing to adapt our technology to integrate with commercial fertilizer-irrigation systems, or fertigation systems, as it’s known in the Ag industry has been a real treat for me personally.</p><p>My prototype today:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dO_ELyiOZ4avqoPy44D_-g.jpeg" /></figure><p>Yes, the peristaltic pumps and all of the electronics are housed in the Amazon box they came in. But hey, it works!</p><h3>GreenThumb Open-Source Hardware Series</h3><p><a href="https://blog.greenthumb.io/open-source-smart-cameras-5b6726918a0e">Smart Camera</a></p><p><a href="https://medium.com/greenthumbio/open-source-smart-cameras-5b6726918a0e">Open-Source Smart Cameras</a></p><p><a href="https://blog.greenthumb.io/greenthumb-io-sensor-kit-66410c52afee?1">Sensor Kit</a></p><p><a href="https://medium.com/greenthumbio/greenthumb-io-sensor-kit-66410c52afee">GreenThumb IO Sensor Kit</a></p><p><a href="https://blog.greenthumb.io/autonomous-plant-feeding-open-source-prototype-df73972e570e">Dosing System</a></p><p><a href="https://medium.com/greenthumbio/autonomous-plant-feeding-open-source-prototype-df73972e570e">Autonomous Plant Feeding Open Source Prototype</a></p><p>If you have any questions or are interested in the project or using my open-source packages, feel free to leave a comment or ping me on <a href="https://linkedin.com/in/tonsoffun111">LinkedIn</a> or <a href="https://www.twitter.com/tonsoffun111">Twitter</a>, I will be responsive.</p><p><a href="https://greenthumb.io/">GreenThumb</a> is also seeking early clients, partners, and initial funding. Yes, I did the work and paid for this all on my own.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f88d884f815" width="1" height="1" alt=""><hr><p><a href="https://medium.com/greenthumbio/greenthumb-opensource-hardware-f88d884f815">GreenThumb OpenSource Hardware</a> was originally published in <a href="https://medium.com/greenthumbio">GreenThumbIO</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[GreenThumb IO Platform]]></title>
            <link>https://medium.com/greenthumbio/greenthumb-io-platform-d6d09ca7fafb?source=rss-5e886beaba05------2</link>
            <guid isPermaLink="false">https://medium.com/p/d6d09ca7fafb</guid>
            <category><![CDATA[cannabis]]></category>
            <category><![CDATA[greenthumb-io]]></category>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[computer-vision]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Justin Bowen]]></dc:creator>
            <pubDate>Sat, 16 Feb 2019 19:39:20 GMT</pubDate>
            <atom:updated>2019-02-16T19:39:20.138Z</atom:updated>
            <content:encoded><![CDATA[<h4>Cannabis Computer Vision, AI, and Harvester Marketplace</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/985/1*CZChskWYsreE0Fo8BJh5_w.png" /></figure><p><a href="http://www.greenthumb.io/">GreenThumb IO</a> is dedicated to delivering high tech solutions to the cannabis industry in close collaboration with leading industry partners.</p><h4>Today: Plant Health &amp; Growth Data Using AI</h4><p>In the GreenThumb product today, smart cameras monitor cannabis growth to provide real-time data, which is analyzed with our machine learning algorithms to determine plant health, tracking the performance indicators of Leaf Area and Bud Area:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FVd-mexs6f2E%3Ffeature%3Doembed&amp;url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DVd-mexs6f2E&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FVd-mexs6f2E%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="640" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/226bba3f6fa1a5eaca9c8da9ac0162ae/href">https://medium.com/media/226bba3f6fa1a5eaca9c8da9ac0162ae/href</a></iframe><p>In the video above we have graphed 720 minutes at 20 FPS, visualizing a decay in the KPIs of Leaf Area Index and Bud Area Index. Based on this analysis, a grower would be alerted of the deficiency and offered a recommended action such as watering or fertilizing the plants, which can be triggered remotely.</p><h4>Introducing: Harvester App</h4><p>GreenThumb‘s second product offering will be the Harvester App, an app that not only assists growers in tracking their grows and the health of their plants, but will allow industry operators to obtain services, bid for product sales, compete for RFPs (Request for Proposals), and much more.</p><p>GreenThumb will provide integrations with existing systems that cannabis companies already know and love, while providing the tooling necessary for new vendors to package their offerings on the platform.</p><h4>Harvester Marketplace</h4><p>Key to our vision is the Harvester Marketplace. By using the GreenThumb Computer Vision product, cultivators will be able to market and sell their harvests in the Harvester Marketplace at no cost. This will create a truly seamless experience for growers to automatically list and sell their product alongside robust data to provide buyers with proof of quality and compliance, streamlining processes and creating greater trust in buyer and seller relationships.</p><p>This harvest data will include operational records such as interactive time-lapse videos, plant health performance metrics, and video captures of human interactions. Growers will also be able to find lower rates on insurance through our platform by proving maintenance and compliance.</p><h4>Partnerships</h4><p>With Harvester, GreenThumb will develop a network of operators and will allow users to connect to data provider adapters built to integrate with their existing Point of Sales, Seed-to-Sale, Security, Distribution, Sensor Systems, Lab Service, Insurance providers, and accounts, all in a cohesive environment. This comprehensive system will offer service providers new opportunities to target the right audience and grow their businesses.</p><h4>Open Source Initiative</h4><p>GreenThumb will maintain open source projects that include the Harvester app, Computer Vision components, and the hardware for Smart Camera, Sensor, and Dosing kits. This will benefit the platform by facilitating more collaboration and healthy competition in the market. Companies using the packages in good faith and in compliance with open source licensing will be welcomed as innovators in the GreenThumb movement.</p><h4>Seeding Future Growth</h4><p>As a fully bootstrapped endeavor, GreenThumb is now open to initial seed funding in order to build a team and accelerate growth. Outside of the goals stated above, we ultimately aim to fully automate the cannabis cultivation process within the next three years, <strong>becoming the premiere technical provider and source of innovation in the $146B cannabis market globally.</strong></p><p>Please contact justin@greenthumb.io to get involved.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d6d09ca7fafb" width="1" height="1" alt=""><hr><p><a href="https://medium.com/greenthumbio/greenthumb-io-platform-d6d09ca7fafb">GreenThumb IO Platform</a> was originally published in <a href="https://medium.com/greenthumbio">GreenThumbIO</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Cannabis Bud Area]]></title>
            <link>https://medium.com/greenthumbio/cannabis-bud-area-5ef7433e6f45?source=rss-5e886beaba05------2</link>
            <guid isPermaLink="false">https://medium.com/p/5ef7433e6f45</guid>
            <category><![CDATA[opencv]]></category>
            <dc:creator><![CDATA[Justin Bowen]]></dc:creator>
            <pubDate>Wed, 30 May 2018 20:35:31 GMT</pubDate>
            <atom:updated>2023-10-18T21:07:17.734Z</atom:updated>
            <content:encoded><![CDATA[<h4>Baby Steps to Tokenized Cannabis</h4><p><a href="https://tonsoffun.github.io/greenthumb/2018/05/22/greenthumb-bud-area/">https://tonsoffun.github.io/greenthumb/2018/05/22/greenthumb-bud-area/</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7Oh8aPqyFY1AYdmmJcC7bw.jpeg" /><figcaption>Cannabis Computer Vision</figcaption></figure><p>I gleamed my inspiration from <a href="https://www.pyimagesearch.com/2016/10/31/detecting-multiple-bright-spots-in-an-image-with-python-and-opencv/">PyImageSearch.com</a> yet again. I’m basically counting bright <em>green</em> spots, by combing some of the tricks from the last post on <a href="https://medium.com/@jusbowen/cannabis-leaf-area-analysis-d6449dbda07f">Leaf Area Analysis</a>.</p><p><a href="https://www.pyimagesearch.com/2016/10/31/detecting-multiple-bright-spots-in-an-image-with-python-and-opencv/">Detecting multiple bright spots in an image with Python and OpenCV - PyImageSearch</a></p><h4>Dependencies</h4><p>In this <a href="https://hub.docker.com/r/tonsoffun/tensorflow-notebook/">docker image</a> I have all of the dependencies for both Leaf Area and Bud Area including working copies and works in progress. We’re going to be using the same setup as with the Leaf Area notebook. <a href="https://opencv.org/">OpenCV</a> will be used for preprocessing the image and pixel selection by range. <a href="https://matplotlib.org/">Matplotlib</a> is used from rendering images and graphs. We take advantage of <a href="http://scikit-image.org/docs/dev/api/skimage.measure.html#skimage.measure.label">skimage.measure.label</a> for taking measurements and label them. We take these components and use <a href="https://opencv.org/">OpenCV</a> to find and annotate the contours.</p><h4>HSV Range Detector</h4><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/a2bbb7381db35175bc4a0e21fd27451b/href">https://medium.com/media/a2bbb7381db35175bc4a0e21fd27451b/href</a></iframe><p>I pulled the first few lines of the script below from this <a href="https://fredrikaverpil.github.io/2016/07/31/docker-for-mac-and-gui-applications/#comment-3684106407">comment</a>. It was need to run the <a href="https://github.com/jrosebr1/imutils/blob/master/bin/range-detector">range_detector</a> used in this and the <a href="https://medium.com/@jusbowen/cannabis-leaf-area-analysis-d6449dbda07f">previous blog post</a>.</p><p><a href="https://sourabhbajaj.com/blog/2017/02/07/gui-applications-docker-mac/">Running GUI applications using Docker for Mac</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Oh0cwcXTqkSwsoWnSAvnPg.gif" /><figcaption>You can checkout this gif to get an idea of how I go about defining the HSV boundary values.</figcaption></figure><h4>Creating the initial HSV Mask</h4><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/5f7d2e7eabc660db7f42f7070e866189/href">https://medium.com/media/5f7d2e7eabc660db7f42f7070e866189/href</a></iframe><h4>Preview the mask</h4><p>We can easily mask the original image with OpenCV’s <a href="https://docs.opencv.org/3.3.0/d2/de8/group__core__array.html#ga60b4d04b251ba5eb1392c34425497e14">bitwise_and</a> which takes the original image and the mask and outputs a masked image.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/51cc314d7b61f01efc9f27432ebc4054/href">https://medium.com/media/51cc314d7b61f01efc9f27432ebc4054/href</a></iframe><h4>Cannabis Computer Vision</h4><p>This is a side-by-side view of the HSV colorspace which the computer uses to isolate the buds and the masked preview. We use the <a href="https://docs.scipy.org/doc/numpy/reference/generated/numpy.vstack.html">numpy.vstack</a> to <strong>v</strong>ertically <strong>stack</strong> the images. I chose vstack for the gists here and hstack in the notebook, but it really is a matter of developer preference.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/871d869d4e700e9abf661fb95199a9ef/href">https://medium.com/media/871d869d4e700e9abf661fb95199a9ef/href</a></iframe><h4>Refining the Mask</h4><p>We can refine the mask using the same binary threshold, erode, and dilate technique used on the <strong>Leaf Area</strong>. This will clean up the specs of frost leave from the actual <strong>Bud Area</strong>.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/43a16ac243ae039ced5c015386773d04/href">https://medium.com/media/43a16ac243ae039ced5c015386773d04/href</a></iframe><h4>Before and After Cleaning Up the Contours</h4><p>You can see the <strong>vstack </strong>comparing the contours pre and post cleanup.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*j5aAUIw4jVJBDAJa6xfJ4A.jpeg" /><figcaption>You can still see a few smaller contours that escape this sweep, but they’re technically smaller bud sites below the top layer of canopy. Either way we’ll filter them in the next pass below.</figcaption></figure><h4>Enhance Mask with Labeled Components</h4><p>Using <a href="http://scikit-image.org/docs/dev/api/skimage.measure.html#skimage.measure.label">skimage.measure.label</a> we are further filtering out noise by ignoring blobs smaller than 100px. This means buds have to be this size or greater to get added to the labeled mask.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/eae4965fa7a9bcf342f4d514b6852c65/href">https://medium.com/media/eae4965fa7a9bcf342f4d514b6852c65/href</a></iframe><h4>Label the Buds</h4><p>Now we have what we need to label the bud sites on the preview image.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/cdad984140fd05ab75d2263923cf7874/href">https://medium.com/media/cdad984140fd05ab75d2263923cf7874/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tlrMzrMBYCTLYVW9bFmpvQ.jpeg" /><figcaption>You can see the smaller spots, which actually appear to be smaller bud sites below the canopy are not labeled. We can remove or lower the threshold of 100px to further refine the camera’s calibration, but this is a great result for this stage.</figcaption></figure><h4>A Full Copy of the Notebook Below</h4><p>Below is the full copy of the notebook. I also have a docker image <a href="https://hub.docker.com/r/tonsoffun/tensorflow-notebook/">tonsoffun/tensorflow-notebook</a> which I’ve simply added OpenCV 3.3 via <a href="https://anaconda.org/anaconda/python">anaconda</a>. This docker image supports both Bud and <a href="https://medium.com/@jusbowen/cannabis-leaf-area-analysis-d6449dbda07f">Leaf Area Analysis</a> as well. Keep following along as I show you how we can use our OpenCV masks to train a <a href="https://towardsdatascience.com/building-a-custom-mask-rcnn-model-with-tensorflow-object-detection-952f5b0c7ab4">Mask RCNN</a>.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/bac3c8865986dc55e27e52a65cdc4422/href">https://medium.com/media/bac3c8865986dc55e27e52a65cdc4422/href</a></iframe><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*XnISCjeCHef6koxRWiykOg.jpeg" /><figcaption>top: HSV Original, bottom: Final Mask Before Contour Filtering</figcaption></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5ef7433e6f45" width="1" height="1" alt=""><hr><p><a href="https://medium.com/greenthumbio/cannabis-bud-area-5ef7433e6f45">Cannabis Bud Area</a> was originally published in <a href="https://medium.com/greenthumbio">GreenThumbIO</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Cannabis Leaf Area Analysis]]></title>
            <link>https://medium.com/greenthumbio/cannabis-leaf-area-analysis-d6449dbda07f?source=rss-5e886beaba05------2</link>
            <guid isPermaLink="false">https://medium.com/p/d6449dbda07f</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[data-visualization]]></category>
            <category><![CDATA[internet-of-things]]></category>
            <category><![CDATA[computer-vision]]></category>
            <category><![CDATA[cannabis]]></category>
            <dc:creator><![CDATA[Justin Bowen]]></dc:creator>
            <pubDate>Tue, 22 May 2018 20:16:38 GMT</pubDate>
            <atom:updated>2023-10-18T21:06:26.881Z</atom:updated>
            <content:encoded><![CDATA[<h4>The First Step in Tokenizing Cannabis Crops</h4><p><a href="https://tonsoffun.github.io/greenthumb/2018/05/22/greenthumb-leaf-area/">https://tonsoffun.github.io/greenthumb/2018/05/22/greenthumb-leaf-area</a></p><p>In this post I’m going to show you how to use openCV and some basic HSV color space analysis used to mask out non leaf pixels. We can use the mask to measure the leaf area in pixels. You can use basic trig to find out the pixel to millimeter value using the distance and angle of the field of view. I’ve broken my Jupiter notebook into smaller pieces to overcome the limitations of medium’s support for rendering long notebooks (though it did seem to render fine on mobile, I couldn’t see it properly on my MacBook Pro using Chrome). Anyway, I also have a full version of the <a href="https://gist.github.com/TonsOfFun/9ed35c29de2ef32884b7e5cea4144c2e">notebook</a> available for anyone interested.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/1*IjHOnf9J_k7-kbel4AXEKA.jpeg" /></figure><h4>Dependencies</h4><p>We’re going to utilize the <a href="https://github.com/jrosebr1/imutils"><strong>imutils</strong></a> convenience package to resize the images for processing as well as for finding our HSV calibration values. <a href="https://opencv.org/"><strong>OpenCV</strong></a> will be used for preprocessing the image and pixel selection by range. We use <a href="https://matplotlib.org/"><strong>matplotlib</strong></a> is used from rendering images and graphs.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/192161d591213f9bb74aa33efb7064b1/href">https://medium.com/media/192161d591213f9bb74aa33efb7064b1/href</a></iframe><h4>Leaf Masking</h4><p>This is a basic function that converts the BGR (OpenCV is BGR, not RGB) image to HSV. Using the `leaf_lower` and `leaf_upper` HSV colorspace values defined using the <a href="https://github.com/jrosebr1/imutils/blob/master/bin/range-detector">range-detector</a> helper tool found in <a href="https://github.com/jrosebr1/imutils">imutils</a>. These calibration settings work on both of my NestCams situated under the same type of HPS light. This is basically pulled from this <strong>PyImageSearch</strong> article on tracking a <a href="https://www.pyimagesearch.com/2015/09/14/ball-tracking-with-opencv/">green ball</a>.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/30a1b1421abcf8ad6345aaf89d992fa9/href">https://medium.com/media/30a1b1421abcf8ad6345aaf89d992fa9/href</a></iframe><p>The erosions and dilations are to cleanup any smaller blobs picked up by the HSV range. In the next blogpost we will go over how to do additional filtering by analyzing the contours and their sizes. I’m not going to apply that here since the results of the post processing on the mask seem adequate for this stage. We will be replacing this model with a custom trained TensorFlow maks RCNN model. Checkout this <a href="https://towardsdatascience.com/building-a-custom-mask-rcnn-model-with-tensorflow-object-detection-952f5b0c7ab4">blogpost</a> on making a toy detector that can output a pixel-wise mask.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/3c241a453f4da584a26e41126730c251/href">https://medium.com/media/3c241a453f4da584a26e41126730c251/href</a></iframe><p>You can see the function working on the two sample images, above from week one and below from week 2. They represent the plants at two time periods 5 days apart. Now let’s checkout what insights the data has to offer.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/3f8679863832d8827e9a3520a5d77e11/href">https://medium.com/media/3f8679863832d8827e9a3520a5d77e11/href</a></iframe><h4>Health Reports and Graphs</h4><p>Below you can checkout a leaf growth report. 10 square inches or 30 square centimeters of growth assuming our camera was hung at the correct height and give or take some variance due to inconsistent leaf distances from the lens. This gives a cultivator great insight into the general health of their crop just by graphing growth. You might also notice that I scaled the images down for processing so the results would have to be scaled up proportionately. That’s an exercise for another time as this post is just to get the concept of monitoring and graphing plant growth with smart cameras.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/135843772b5f9979a880e01f19619028/href">https://medium.com/media/135843772b5f9979a880e01f19619028/href</a></iframe><p>We can now have the ability to provide reports on performance as well as alerts when growth has plateaued or even dropped off which is a clear sign of stress in any plant. We can also provide historical performance reviews so a cultivator can see how the current crop is vegging in comparison with previous crops at the same stage of growth.</p><p>Let’s plot these data points from our two leaf area analyses. Since we only have our two data points from this time series our line graph will be a little boring, but you can imagine the growth being more engaging when plotting a point every minute instead of every 5 days. A realtime growth chart would include oscillations from the circadian rhythms induced by the light cycles, as well as fluctuating from variable environmental conditions. The decentralized AI will have to compensate for variability dynamically at each camera.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/d149111ba1f5bb6b68b9dbb8b7ea8ab2/href">https://medium.com/media/d149111ba1f5bb6b68b9dbb8b7ea8ab2/href</a></iframe><p>With this functionality we increase the farmers awareness of problem areas in there grow sites and give them metrics to back their business’s performance overtime. We can now take photos of cannabis plants and produce metrics on our devices. It’s the first step on our path to a decentralized cannabis AI and there is enormous potential value in a trusted data source of cannabis cultivation metrics from seed to harvest. I’ll be putting out an article later this month talking about GreenThumb IO’s plans to empower the farmers with their own crop data with GTIO Security Tokens.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jc0efTe4Ug77dOpmlj4T3A.jpeg" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d6449dbda07f" width="1" height="1" alt=""><hr><p><a href="https://medium.com/greenthumbio/cannabis-leaf-area-analysis-d6449dbda07f">Cannabis Leaf Area Analysis</a> was originally published in <a href="https://medium.com/greenthumbio">GreenThumbIO</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[IoT Smart Camera Landscape]]></title>
            <link>https://medium.datadriveninvestor.com/iot-smart-camera-landscape-942442f88c8?source=rss-5e886beaba05------2</link>
            <guid isPermaLink="false">https://medium.com/p/942442f88c8</guid>
            <category><![CDATA[computer-vision]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[deep-learning]]></category>
            <category><![CDATA[iot]]></category>
            <dc:creator><![CDATA[Justin Bowen]]></dc:creator>
            <pubDate>Thu, 17 May 2018 05:07:59 GMT</pubDate>
            <atom:updated>2019-03-02T15:26:40.406Z</atom:updated>
            <content:encoded><![CDATA[<h4>Qualcomm Vision Intelligence Platform + Microsoft Azure</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/617/1*EFNAReQdtzICBHB7G9M8sA.png" /><figcaption><strong>Vision AI Developer Kit</strong></figcaption></figure><p><a href="https://www.qualcomm.com/news/onq/2018/05/07/qualcomm-vision-intelligence-platform-microsoft-azure-bring-edge-ai-solution?cmpid=oofyus181390">Qualcomm</a> just announced a device aimed at developers in partnership with <strong>Microsoft</strong> and their <strong>Azure</strong> platform. <strong>Microsoft</strong> and <strong>Qualcomm</strong> are coming out with a <a href="https://www.visionaidevkit.com/">Vision AI Developer Kit</a> which you can signup and learn more about now. From what I can see they’re offering a cloud based <a href="https://azure.microsoft.com/en-us/overview/machine-learning/">Azure Machine Learning Service</a> coupled with their <a href="https://azure.microsoft.com/en-us/services/iot-edge/">Azure IoT Edge Platform</a> to take your cloud trained <strong>AI</strong> model and run them on cross platform edge devices, like smart cameras. All of this along with what appears to be a <strong>dockerized</strong> deployment of your cloud trained models to be run on device.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZbeXTl1MsqRZP1USoLb7xg.png" /><figcaption>Qualcomm currently powers the NestCam IQ</figcaption></figure><p>All of it powered by <strong>Qualcomm’s</strong> new <a href="https://www.qualcomm.com/products/qualcomm-vision-intelligence-platform">Visual Intelligence Platform</a> and <a href="https://developer.qualcomm.com/software/snapdragon-neural-processing-engine-ai/getting-started">Snapdragon Neural Processing Engine</a> (NPE-yet another acronym). They have been talking about these types of devices since since <a href="https://www.theverge.com/2015/10/26/9616172/qualcomm-home-camera-reference-design">2015</a>. The details I can find on the internet are a little fuzzy on to whether the <a href="https://www.theverge.com/circuitbreaker/2018/4/11/17218686/qualcomm-smart-camera-chips-vision-intelligence-platform">NestCam IQ</a> is powered by one of these newer chips or the older ones repurposed to do <strong>AI</strong> on device. Either way it is great to see other chipset manufacturers and software service providers involved in innovating smart device development.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*7NBFgyJzoj8JZCcr9W9kQQ.png" /><figcaption>Flow diagram provided in the blogpost linked about. Illustrated the pre-canned and cloud trained model support provided by <strong>Microsoft Azure</strong> and <strong>Docker</strong> on the <strong>Qualcomm</strong> powered Smart Camera Device</figcaption></figure><h4>Amazon’s AWS Deep Lens</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/610/1*USs6Y1jeLwJwpFuFy4-Jhw.png" /><figcaption>The AWS approach to IoT</figcaption></figure><p>This is yet another smart camera developer kit to hit the market and it is great news for <a href="https://greenthumb.io/">GreenThumb IO</a>. I’m excited to get my hands on the <a href="https://aws.amazon.com/deeplens/">Amazon Deep Lens</a> which will be release June 14, 2018. Amazon has teamed up with Intel and their <a href="https://www.intel.com/content/www/us/en/products/processors/atom/e-series.html">Atomprocessor</a>, which they quote <a href="https://docs.aws.amazon.com/deeplens/latest/dg/what-is-deeplens.html">here</a> can process 100 billion floating-point operations per second (GFLOPS). While it’s not out yet they’re already blogging about it <a href="https://aws.amazon.com/blogs/aws/deeplens/">here</a>. It looks like they’re giving you the power of AWS <a href="https://aws.amazon.com/greengrass/">Greengrass</a> and <a href="https://aws.amazon.com/lambda/">Lambda</a> on device. This is going to enable cloud power AI solutions to break free of the chains of centralization and go to edge.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/861/1*qJ83kZUoYwUaIqHXHwFNsg.jpeg" /><figcaption>Amazon’s take on a developer friendly smart camera</figcaption></figure><blockquote>it is safe to say that <a href="https://aws.amazon.com/iot/">IoT</a> and <a href="https://aws.amazon.com/serverless/">serverless</a> computing are here to stay and should be on your list. — <a href="https://aws.amazon.com/blogs/aws/deeplens/">https://aws.amazon.com/blogs/aws/deeplens/</a></blockquote><h4>Google’s AIY Vision Kit</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/640/1*O1pHnR3VWzr_sCn1WQSsDg.png" /><figcaption>The Movidius VPU Chip</figcaption></figure><p>I’ve mentioned on my first <a href="https://medium.com/p/c878a8416cb5?source=linkShare-5e886beaba05-1526523881">blog post</a> that the Alphabet is ahead of the curve again with the <a href="https://aiyprojects.withgoogle.com/vision">Google AIY Vision Kit</a>. Launched November 30, 2017, the AIY Vision Kit piggybacks off an <strong>ARM</strong> <strong>CPU</strong> powered <strong>Raspberry Pi</strong>, with a vision bonnet that features a <a href="https://www.movidius.com/news/google-launches-aiy-vision-kit-featuring-intel-movidius-vpu">Movidius Chipset</a>. This <strong>VPU</strong> coprocessor is used to execute TensorFlow models on the device while the <strong>ARM CPU</strong> is used for program logic. It’s encouraging to see so many big companies enabling developers with these proprietary devices with fully programmable with open source software.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JZOiQpxRB2V0XnQuzy3Hlg.jpeg" /><figcaption>Google’s AIY Vision Kit is the only Camera based AI Vision <strong>developer kit</strong> in <a href="https://www.target.com/p/-/A-53417081">stores now</a></figcaption></figure><p>It would appear the giants are gearing up for next revolution of smart devices. It seems to be a competition for operating platforms and services. As someone with aspirations of my own it makes me wonder which one of these companies is ready to start partnering with startups to encourage growth in this area.</p><h4>NVIDIA’s Jetson Platform</h4><blockquote>the embedded platform for autonomous everything. — <a href="https://www.nvidia.com/en-us/autonomous-machines/embedded-systems-dev-kits-modules/">NVIDIA on Jetson</a></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/625/1*VSu_a445CjCK4hE2-fIJSA.jpeg" /></figure><p>NVIDIA, which I mention in my <a href="https://medium.com/p/3a512c75acf2?source=linkShare-5e886beaba05-1526524168">last blog</a>, also has a <strong>developer kit</strong> of their own with the <a href="https://www.nvidia.com/en-us/autonomous-machines/embedded-systems-dev-kits-modules/">Jetson Platform</a>, the embedded platform for autonomous everything. Its powered by the awesome <strong>Tegra 4.0</strong> chipset which boast a beefy <strong>ARM CPU</strong> as well as<strong> 256 CUDA</strong> enabled <strong>GPU</strong> cores. <strong>NVIDIA</strong> has been powering both the crypto and artificial intelligence hypes and it’s been reportedly taking a toll on their stocks. Regardless <strong>NVIDIA</strong> is the clear leader in <strong>GPU/TPU</strong> technology as well as open source software. <strong>CUDA</strong> has paved the way for modern day deep learning frameworks such as <strong>TensorFlow</strong>, <strong>Caffe</strong>, <strong>Theano</strong>, <strong>PyTorch</strong>, and so on.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/625/1*0bzKSBT2GpuoNGqjF_TIyg.jpeg" /><figcaption>NVIDIA offers a barebones <strong>Developer Kit</strong> which you can plug-n-play a number of camera, audio, and other peripheral devices into. This platform is what I aim to use in the next phase of research &amp; development for GreenThumb IO’s CryptoSmartCamera Platform</figcaption></figure><blockquote>Where do you end up once you have a marketable smart camera product? — me just now!</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/500/0*NzxnrSpmT3zMVY_4.jpg" /><figcaption>Ideally we could deploy our code directly to a standard IP66 Power Over Ethernet security camera!</figcaption></figure><p>So what does it mean for the ambitious developers and entrepreneurs out there. Personally i think it’s inspiring, but if you build something cool with any one of these platforms you have to ask yourself a question. Where do you end up once you have a marketable smart camera product? I’ve submitted an application for <a href="https://greenthumb.io/">GreenThumb IO</a> to join the <a href="https://www.nvidia.com/en-us/deep-learning-ai/startups/">NVIDIA inception program</a>, you can read my application below. Enjoy!</p><blockquote>GreenThumb IO is developing a Smart Camera platform for commercial cannabis cultivators using Blockchain-powered decentralized Artificial Intelligence.</blockquote><blockquote>Utilizing over 1000 cameras per acre, powered by the NVIDIA Tegra platform, we will actively monitor and learn at the commercial scale of hundreds of thousands acres. We will use Deep Learning techniques to not only enhance the Computer Vision that powers the analytics, but also cultivate an artificially intelligent agronomic understanding of key visual growth indicators.</blockquote><blockquote>We’re starting out with TensorFlow trained models that can infer and mask leaf area, bud area, stress, and harvestability at a millimeter per pixel resolution. These enormous high resolution datasets, coupled with the power of NVIDIA’s distributed CUDA framework and GreenThumb IO’s distributed AI, will create an unprecedented super human understanding of cannabis agronomy.</blockquote><blockquote>The system is designed to transmit metrics from the Computer Vision Models to blocks. The cameras GPUs participate in solving blocks when they’re not capturing and analyzing their FOV. We hope to run the TensorFlow inference on the ARM CPU. This is a practical and scalable blockchain solution that will limit if not virtually eliminate cloud servers.</blockquote><blockquote>With NVIDIA, GreenThumb IO will be the most advanced Decentralized AI Smart Camera Platform.</blockquote><blockquote>Justin Bowen, GreenThumb IO</blockquote><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=942442f88c8" width="1" height="1" alt=""><hr><p><a href="https://medium.datadriveninvestor.com/iot-smart-camera-landscape-942442f88c8">IoT Smart Camera Landscape</a> was originally published in <a href="https://medium.datadriveninvestor.com/">DataDrivenInvestor</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>