This repository attempts to build a lane detection system using both OpenCV and deep learning techniques with the goal of deploying onto an Autonomous Mobile Robot (AMR) via a Raspberry Pi. Currently, this repo contains the OpenCV methods adapted from this GitHub solution.
- Grayscale the image
- Apply Gaussian blur
- Use Canny edge detection
- Define vertices for area of interest and apply mask
- Use Hough Lines Transform to find lines
- Merge detected lines and final image
- Create pipeline
- OpenCV
- Matplotlib
- Numpy
- To use deep learning techniques to train a model for lane detection instead of using image processing methods.
- To attempt on different embedded systems like Jetson Nano for deployment.