Abstract
We propose a method for detecting vehicles from an image.At first ,two edge images are used for deciding candidate area.In the area of grayscale image,using Gabor filters that can extract frequency elements of any direction,the intensity of an input image is transformed into Gabor energies which has information about spatial extent and direction of the objects.Gabor energies and color information of an input image are used for segmentation of the image.And using Affine Moment Invariants,vehicles are detected with the information of the segmented image.
