In this project,the drivable area for an experimental robot car (powered by Raspberry Pi 4 and motor driver) has been detected using Image Segmentation and the binary output mask is obtained. The midpoints of the row-wise maximum width sections of the binary output mask have been joined to create a path which the car can follow. This prediction keeps running and changes dynamically according to the image stream recieved from the robot car.
- Opencv-python
- tensorflow
- pixellib
Download model weights from here