Performance Evaluation of Vehicle Detection Methods

The vehicle detection is the process of marking the vehicles in the images. The vehicle detection processes can be performed on the aerial, satellite, vehicular camera or other similar or non-similar images. The vehicle detection methods are used for vehicle number or vehicle density calculations, vehicle tracking, vehicle monitoring, urban vehicular activity monitoring, etc. Such application requires a high accuracy to perform their operations accurately. For the purpose of higher vehicle detection accuracy, there are several techniques proposed by various authors. In this survey, the authors have proposed the solution using template matching, self learning, feature extraction, feature analysis or various other techniques to detect the vehicles in the aerial, satellite or other images. The existing techniques have been evaluated for their adaptability and accessibility. Also the existing techniques have been analyzed on the basis of various common performance parameters. The image processing techniques based on template matching methods have been found performing better than the other image processing techniques used for the vehicle detection.