There’s no doubt that advances in vision technology (algorithms, data, computation and software) are transforming the way fields in agriculture are managed. The unit of management is moving from whole fields and farms to individual plants and even insects. Yet the approach taken—generally closed source, proprietary solutions—mean we aren’t innovating as fast as we could, and we have removed the tools for innovation from the key innovators themselves: farmers. Using his experience with the open-source OpenWeedLocator for DIY weed recognition and the WeedAI platform for image data sharing, Guy will discuss how transforming our approach to an open-source model, will benefit companies, farmers, and the future of food production.
#computervision #ai #artificialintelligence #machinevision #machinelearning #datascience #opensource
Contents of This Video --
00:00 - Introduction
01:10 - Defining Agriculture
02:22 - Edge Cases in Agriculture
04:07 - Technology and Field Observation Intensity
06:01 - Deep Learning and Scalability Challenges
06:52 - Weed Detection Techniques
09:19 - Vision AI in Practice
11:06 - Rise of Deep Learning in Agriculture
13:10 - Variability and Limitations of AI Models
15:01 - Barriers to Access and Innovation
16:33 - The Power of Open Source in Agriculture
19:03 - Community and Collaboration Through Open Source
21:31 - The Open Weed Locator
26:36 - The Future of the Open Weed Locator
27:14 - Open Source Image Data and Sharing Platforms
28:54 - Final Thoughts and Call for Openness
30:16 - Audience Q&A