Vehicle Detection and Classification in Low Vision Video Footage

This proposed research work describes a system that is used for recognition and classification of vehicles from low vision video footage captured by surveillance cameras. It's an important step for many applications in intelligent transportation systems and video surveillance. For carrying out the above said process, first step is detection of the vehicles present in the enhanced video frame & then classification of a particular vehicle based on vehicle features like color, shape and corner points in the video. An innovative method is proposed for recognizing the vehicle present in the video. At first, enhancement of the low/night vision footage is done. After it, the features are extracted. Color is extracted using RGB color values of the vehicle, Shape is extracted based upon Zernike moments & for Corner detection, Harris corner detector process is applied. Finally the Fuzzy Inference system takes the help of extracted features to classify the type of vehicle.