A Resolution Of Video Image Stabilization Based On Gray-coded Bit-planes Matching

In this paper, a video image stabilization algorithm, which was based on Gray-Coded Bit-Planes matching, was investigated and realized. The algorithm can reduce the irregular motion of the real video signal and smooth the video sequences. Experimental results showed that the algorithms exhibited good performance when applied to the horizon and vertical motion and significantly reduced computational efficiency, convenient for the development of hardware.The first part of the paper, the basic theory of the algorithm was introduced. Post-processing the image and the Gray-coded Bit-planes were achieved, based on the minimum correlation measure to search the most matching image through three step searching algorithm and estimate the motion parameters between the previous and current image. The compensation parameters were generated through respective weights. Then the stable image sequence can be obtained by compensating image frames with the compensation parameters. Visual C++ 6.0 was applied to realize some simple graphic interface, and the finished functions were as follows: read the AVI files and transformed these files to gray-coded bit-planes, bit planes matched and estimated the motion parameters, motion compensated, writ AVI files and play AVI files.The general work of this paper was summarized and the further research direction was pointed out in the last of the paper.