Comparison Research On The Compression Algorithms In Low Bit Rate For Aerial Video Images

With the rapid development of science and technology, aerial photography is widely applied in civil and military domains. Due to the large amount of data containing in the video images, the way of compressing aerial video images through low bit rate compression algorithm is adopted to solve the problem of bandwidth in the aerial wireless telecommunications and storage. Thus, it seems important to find a kind of efficient and low bit rate compression algorithm which fits for aerial video images.By comparing with ground sampled distance and interframe motion vectors of aerial video images, it has the characteristic of the low ground sampled distance, slow interframe motion and consistency of motion field. Based on the characteristic of aerial video images and conditions of low bit rate video applications, the low bit rate compression algorithm is aptly selected to compress aerial video images. Considering the algorithm performance, complexity and application, H.261, H.263, MPEG-4 and H.264 algorithm are compared and H.263 algorithm is finally selected. H.263 algorithm is realized with VC++6.0 compile environment, large numbers of experiment on aerial video images are processed and the performance of H.263 algorithm is compared under different parameters. Motion estimation algorithm based on blocks was deeply researched, it is improved in half-pixel motion estimation algorithm, improved full search algorithm and predicting interframe motion vector with motion vector of individual macro block, the improved motion estimation algorithm based on aerial video images is given out in line with the characteristic of aerial video images.In this paper, the innovation is that consistency of motion field of aerial video images is utilized to improve motion estimation algorithm. The advantage of the improved motion estimation algorithm is that the operational complexity of motion estimating algorithm is reduced. By the framework of H.263 algorithm and improved motion estimation algorithm, a framework of H.263 algorithm fit for aerial video images is given out. This algorithm has a high application value which can be adopted in aerial photography.