Image Compression And Stream Medium For Traffic Video Based On Wireless Network

This paper aimed at the characteristic of the big deal of data in image transfersand the characteristic of the traffic image itself. And we will research the real trafficimage compression and transfers.The general content of this paper is brief reviewing of the history on datacompression, achieving the research of real traffic image compression arithmetic andthe correlative base arithmetic. And also imagining the landscape of the wirelesstransfers of compressed data. Main contents include the research on imageclassification and the application on real traffic image; the grey image fractalcompression and the color image fractal compression coding using the correlation ofcolor components; high speed polynomial approaching coding method; approachingcoding method based on GM (1,1); the research on stability measurement for imagescanning data suitable for the idea of approaching; the research on self-adaptparameter choice technique for polynomial approaching; and the research on Huffmancoding based on number character. At last, this paper introduces the Real System SDKsimply and actualizing of the stream medium for traffic video based on wirelessnetwork.The research show fractal dimension can be a good parameter in image classifiedand analysis. The classified result based on fractal dimension is the same as the resultof classified according to person's subjective feeling. As for traffic image, thedistributing zone of fractal dimension respectively is 2.4-3.0 and 2.8-3.0, when box'sside is 2 and 4. The fractal method is of high coding and decoding rates, and withquite good signal to noise ratio, compression rate and visual quality about almostevery image. For some image, this method exceeds the JPEG. The method based onthe idea of approach show the PSNR is bigger than 20, while the compression rate issmaller than 20% and the majority summation of the positive value and the negativevalue in the error image are all below 150 to a small complex traffic image which is256 multiply 256. The stability measurement for image scanning data can not onlyevaluate the known method of image scanning, but also guide the design for themethod of image scanning. Huffman coding method based on number characterreduce the traditional Huffman table in some extent while enhance the compressionratio of the image data. Experiment data indicate that the Huffman table can becontrolled below 40bit, while the coding efficiency may amount to 95% above, thecompression ratio can amount to 60% above.