This repository contains my work on a couple of computer vision tasks related to autonomous driving and perception. The projects mainly use OpenCV and NumPy to process images and video and detect useful road features.
This folder contains my solution to a perception coding challenge.
The goal here was to detect red regions in an image and outline them in a useful way. My approach was to:
- convert the image from BGR to HSV
- isolate the red regions using color thresholds
- apply the mask to the image
- convert the result to grayscale and threshold it
- find contours
- split the image into left and right halves
- connect contour points to mark the detected regions
Files:
wisco.py— main scriptanswer.png— output imageREADME.md— explanation of the approach
This folder contains my work on detecting stop lines from road video frames.
The script processes a video frame by frame and tries to detect horizontal stop lines by:
- selecting a region of interest from the road
- converting the region to HSV
- isolating bright white areas
- filtering by intensity
- applying morphological operations
- detecting edges with Canny
- detecting lines with Hough Transform
- keeping only near-horizontal lines
Files:
stoplinedet.py— main scriptREADME.md— explanation of the approach
- Python
- OpenCV
- NumPy
- math
Install dependencies:
pip install opencv-python numpyThen run the scripts from their folders:
python wisco.pypython stoplinedet.pyNote: some file paths in the scripts may need to be updated to match your local machine.
Overall, this repository shows my work in building and testing classical computer vision pipelines for perception-related problems. It helped me get more comfortable with image preprocessing, masking, contour detection, edge detection, and line detection in OpenCV.