justincinmd / carnd-vehicle-detection

Vehicle Detection Project

Vehicle Detection

Submission Notes

To detect cars, I:

• Perform a Histogram of Oriented Gradients (HOG) feature extraction on a labeled training set of images and train a Linear SVM classifier
• Append a histograms of color to my HOG feature vector.
• Normalize my features and randomize a selection for training and testing.
• Implement a sliding-window technique and use my trained classifier to search for vehicles in images.
• Run my pipeline on a video stream and create a heat map of recurring detections frame by frame to reject outliers and follow detected vehicles.
• Estimate a bounding box for vehicles detected.

See Vehicle Tracking.ipynb file for a complete overview of the pipeline, with example images and a more extensive overview.

The output.mp4 file shows the final detected cars, and the debug_output.mp4 file contains the output image, as well as images from each stage of the pipeline.