The real-time image processing technique based on DSP

Abstract

This paper proposes a novel real-time image processing technique based on digital singnal processor (DSP). At the aspect of wavelet transform(WT) algorithm, the technique uses algorithm of second generation wavelet transform-lifting scheme WT that has low calculation complexity property for the 2-D image data processing. Since the processing effect of lifting scheme WT for 1-D data is better than the effect of it for 2-D data obviously, this paper proposes a reformative processing method: Transform 2-D image data to 1-D data sequence by linearization method, then process the 1-D data sequence by algorithm of lifting scheme WT. The method changes the image convolution mode, which based on the cross filtering of rows and colums. At the aspect of hardware realization, the technique optimizes the program structure of DSP to exert the operation power with the in-chip memorizer of DSP. The experiment results show that the real-time image processing technique proposed in this paper can meet the real-time requirement of video-image transmitting in the video surveillance system of electric power. So the technique is a feasible and efficient DSP solution.