For FRC Deep Space Tensorflow Object Detection. Used a modified tiny-yolo (v2) configuration from darknet for training weights. Hatch.py was modified from AlexeyAB's darknet fork. For demonstation purposes, opencv shows webcam feed, however can be removed for FRC deployment. Using pynetworktables, whether a box is detected is transmitted as well as the coordinates of drawn bounding boxes. For other frc teams, note the following change needed, change the 555 on line 10 to match your team number.
Dependecies:
- Darknet
- Opencv
- Pynetworktables