An Improved Coding Mechanism For Bayer Pattern Images

In most of the compressed encoding scheme, we first do separation and transformation of structure to the three primary components of the Bayer image and remove redundant through transformation, DPCM, quantization etc.Finally, the statistical encoding is carried. Due to differences of specific technical route, each decoding performance has advantages and disadvantages. It is not hard to find that direct compression mechanism of organic combination of transform and quantization can get higher rate-distortion performance. But the algorithm also has some problems, such as higher rate-distortion performance at the expense of the higher complexity and higher coding complexity than decoding end. On the other hand, Each separated component is compressed alone by JPEG algorithm, and not for global optimization. So these algorithms posses some problems, such as each separated component quantizer without the global optimization. In the process of communication, the rate-distortion performance can be improved further, because channel coding and entropy coding work independently. Therefore, it is necessary to discover a new Bayer pattern images method.First, in order to solve the problems of low image lossless compression, the poor quality of reconstructed images and the high complexity, this paper mainly considers the quantization in the encoding end. We emphatically consider WZ quantizer’s optimal conditions under the condition of joint entropy constraint. And a Lloyd iteration algorithm is proposed on the base of the convergence of Lagrange cost function. Furthermore we design a global optimized quantizer and do simulation experiments to prove that this algorithm is better than the traditional algorithm in guaranteeing the quantizer local optimal ity.Second, this paper makes a research on the characteristics of Bayer pattern, image coding theory and S-W coding.Then we mainly adopts the RCPT code, which adds a punch module in the front of Turbo code, and haves a better performance.What’s more, this paper use the DCT on the orthogonal transformation,.Finally, in order to solve the existing problems in Bayer pattern for digital image decoding method such as global optimization, the improvement of rate-distortion performance, it provides a Bayer pattern based on the structure of W-Z digital image decoding method, which does separation and transformation of structure in the encoding end and forms a four component image and performs discrete cosine transform respectively. A global optimized quantizer is designed to the convergence of Lagrange cost function and Lloyd algorithm. Quantitative output for each component of the transform coefficient is encoded independently based on w-z coding method. The decoding effectively uses of luminance component as side information and jointly decodes and reconstructs bill template image.The experimental results show that this method can decrease the encoding complexity and get better improvement of the rate-distortion performance under the condition of high speed. But the quantization noise increase leads to the system SNR fast falling at a low rate. Therefore, we can conclude that this improvement is reasonable, which solves the existing problems of the less global optimization and the poor rate-distortion performance in Bayer pattern coding of quantizer design.