Project 5: Vehicle Detection and Tracking

Overview

Identify vehicles in a video from a front-facing camera on a car using image classifiers such as SVMs and HOG.

The goals / steps of this project are the following:

• Perform a Histogram of Oriented Gradients (HOG) feature extraction on a labeled training set of images and train a classifier Linear SVM classifier
• Apply a color transform and append binned color features, as well as histograms of color, to the HOG feature vector.
• Implement a sliding-window technique and use the trained classifier to search for vehicles in images.
• Run the 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.

Project Deliverables

• train.py to train the Linear SVM model with set parameters. The training data and parameters are saved in a pickle file for later use.
• extract.py to extract features using hog sub-sampling and make predictions.
• box.py to store windows found over a set number of frames, get heatmap, and return image with final bounding box
• video.py to produce the video with bounding boxes and number of detected vehicles displayed.
• project_video_output.mp4 is the final video with smoother detection by accounting for previous frames

Results

View the video on Youtube