Research On Interframe Compression Of Bayer Sequential Images

Since the 3CCD professional camera uses three sensors(3CCD)to obtain the color values of three channels of each pixel of the image,and then obtain the full-color image,the 3CCD professional camera is expensive and difficult to manufacture.Currently,digital color cameras generally use a single sensor with a color filter array(CFA)to capture images,and Bayer arrays are the most widely used.With the rapid increase of image data in Bayer format,the channel bandwidth and data storage capacity are greatly stressed,and image compression is the most effective method to solve this problem.Since each pixel of Bayer array can only obtain the color information of one channel,the lack of information makes the Bayer format image unsuitable for direct compression,so it needs to be transformed into an image suitable for compression through the image transformation process.The "Structure Conversion" method is the most widely used image transformation algorithm,but the low-pass filter removes the high frequency components caused by the loss of pixels while also removing the high frequency of the image edge,which reduces the quality of the reconstructed image.In order to avoid blurring out some image edges with low-pass filtering,this paper proposes to use Sobel edge detection operator to enhance the edges after the low-pass filtering algorithm,so as to reduce the distortion at the edges of the image and improve the quality of the reconstructed image.Since the previous Bayer image compression algorithms seldom consider time redundancy,this paper introduces the interframe compression algorithm into the Bayer image compression technology.In this paper,a patch matching motion estimation algorithm based on fireworks algorithm is proposed for motion compensation module of interframe compression.Compared with the traditional patch matching algorithm,this algorithm shows excellent performance and high efficiency.Different image regions have different characteristics,and the regions with abundant details and edges need higher reconstruction quality.This paper proposes an adaptive image transform algorithm based on region.As the region of interest of the image,the local moving patch detected by the motion estimation module often has rich edges and details.The "Structure Conversion" algorithm based on edge enhancement with better reconstruction quality is elected for the local moving patches,while the "Structure Conversion" algorithm with lower complexity is selected for the other patches.With this adaptive mode,the speed and reconstruction quality can be considered,and finally the efficient compression of Bayer format images can be realized.The experimental results show that the proposed method has higher compression ratio and reconstruction quality than the traditional "Structure Conversion" algorithm,especially in the moving sequence image.