Contour Based Car Recognition Algorithm

This paper proposes a car recognition algorithm based on multiple features of contour. Firstly, canny operator is used to get the edge of the image. Afterwards, image pyramid is used to shrink the image so that the contour will be single-edged and complete. Then, from a point in the contour, travel the whole contour to gain the Fourier descriptors and direction ratio from the traversal sequence. At last, all features are considered together to implement the car recognition. The experimental results of many car images show that our method has fine accuracy in car recognition.