https://github.com/montclairrobotics/deepspaceml
For FRC Deep Space Tensorflow Object Detection
https://github.com/montclairrobotics/deepspaceml
Last synced: 3 months ago
JSON representation
For FRC Deep Space Tensorflow Object Detection
- Host: GitHub
- URL: https://github.com/montclairrobotics/deepspaceml
- Owner: MontclairRobotics
- License: other
- Created: 2019-01-09T02:45:52.000Z (about 7 years ago)
- Default Branch: master
- Last Pushed: 2019-02-04T15:06:26.000Z (about 7 years ago)
- Last Synced: 2025-03-11T15:43:56.742Z (about 1 year ago)
- Language: Python
- Size: 575 MB
- Stars: 0
- Watchers: 1
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# DeepSpaceML
For FRC Deep Space Tensorflow Object Detection. Used a modified tiny-yolo (v2) configuration from [darknet](https://pjreddie.com/darknet/) for training weights. Hatch.py was modified from [AlexeyAB's darknet fork](https://github.com/AlexeyAB/darknet). For demonstation purposes, opencv shows webcam feed, however can be removed for FRC deployment. Using [pynetworktables](https://robotpy.readthedocs.io/en/latest/install/pynetworktables.html), 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