OpenCV is the most important library for computer vision, and as a way to celebrate its 20 years of creation, they decided to launch the OpenCV Spatial AI Competition, sponsored by Intel and with its strategic partner Luxonis. The competition consisted in solving real world problems with the OAK-D (OpenCV AI Kit with Depth) module that combines the power of neural inference with depth perception in real time. So we decided to submit our proposal to monitor weeds, detect their species and measure their biomass, in order to generate much more efficient control mechanisms. Our proposal was selected as a strong input and this is how we have spent the last three months adjusting and developing AI methods to achieve our goal.
Please read our wiki to see our progress. https://github.com/precision-sustainable-ag/PhenoCV-WeedCam/wiki
We use this repository privately initially to develop the code and configure the RPi platform to detect Weeds using the OAK-D camera. Here is the description of what we want to achieve after we get proper results.
https://docs.google.com/document/d/18wOEV-NvKzdh4ZtCB1QlT4qAeVp_5Y2HzEcZbvxXgXc/edit
We will work under a spring backlog methodology where we will perform springs every week, for four weeks, trying to reach our goal. We will also have the support of Luxonis, and Intel.
https://trello.com/b/Y2AjVqnt/oak-d-competition
Other resources