Vision Based Relative Speed Detection of Moving Objects in a Traffic Environment

Speed estimation of an object from a camera is a challenging task in the field of computer vision. Speed of the moving object from a moving camera is furthermore challenging. In this paper, a camera is used to acquire the images. The edges are detected and broken edges are merged to connect the broken components. The corners are detected to further find the optical flow of the same. The points are clustered based on the direction of optical flow of each object, thus marking the same as moving object. The object detected is used to find the displacement with a novel method of view metrology. Further, this is used to find the relative speed of the object, based on the frame rate of the camera used. The errors are computed with the known results with constant speed limits fixed initially.