Keywords: Linear Support Vector Classifier, Feature Extraction, Histogram of Oriented Gradients (HOG), OpenCV
Goal of this project is to write a software pipeline to identify vehicles in a video from a front-facing camera on a car.
- HOG features are extracted from car and non-car images and a data set is constructed.
- A linear SVC is trained, on HOG features, to classify car and non-car images.
- A sliding window search is used and a heat map is constructed from it to identify cars in front facing carmera image.
- Ipython notebook is here, detailed project writeup is here and final video output is here.