Improvement Research Of Color Interpolation Algorithm And Circuit Design

Digital Camera technology is developing rapidly and popular in recent years. The color interpolation technology is an important part of the internal process of the digital camera. Color interpolation technique so far has been two constraints. One is high-performance. The error of interpolated image should be as small as possible. The other is low-cost. The design should use as few resources as possible. In this paper, two improved algorithms are proposed according to surviving low-cost algorithm and high performance algorithm. The improved high performance algorithm is also verified in hardware.Typical algorithms which can be realized by hardware are compared and analyzed and two of them are selected as main study objects. One is the low-cost Hibbard edge detection algorithm. Sobel operator is used to improve the gradient detection and reduce interpolation error. The improved algorithm increases the average PSNR on R, G, and B channels by 0.41db,0.45db and 0.4db respectively. The other is high-performance Hamilton & Adams adaptive interpolation algorithm. It is proposed to improve gradient detection using color difference information. The improved algorithm can effectively inhibit false color appears on dense lines of images. It increases the average PSNR on R, G, and B channels by 0.59db,0.69db and 0.56db respectively.Based on the improved Hamilton & Adams algorithm, hardware circuits of color interpolation is designed and implemented on chip Cyclone Ⅳ EP4CE115 FPGA. The design can run up to 100MHz and meet the requirement of 1080p@30fps.