ABSTRACT
Object tracking in multi-camera systems has been widely used in social life, however object re-identification between cameras is difficult and inefficient. It will improve the efficiency of object re-identification by applying fast and real-time compressive sensing feature into multi-camera. Due to the lighting conditions and projection size of an object in different cameras being different, while a compressive feature is sensitive to the greyscale and size, for the greyscale and size problems, this paper puts forward the idea of unification. It maps the compressive feature value of the sample frame to the value of the original frame by the ratio of the average greyscale value and the ratio of the area of two frames. Experiments show that the algorithm proposed in this paper behaves quickly and accurately in object re-identification when lighting and object size changed.
