基于双芯片DM642的Bayer格式图像修复系统

A Bayer image restoration system based on double chips DM642

提出了一种基于数字信号处理器(DM642)和可编程逻辑门阵列(FPGA)架构的Bayer格式图像修复系统,可应用于高速图像采集和实时处理.该系统采用双处理器流水线体系结构,DM642计算出改进的四点像素Bayer格式图像校正参数,FPGA根据校正后的参数对失去的像素进行恢复,再按一定的数据格式打包传给DM642进行双线性插值运算.实验结果表明:该系统修复一幅大小为752×480像素的Bayer图像大约需要15 ms,完全可以满足实时性的要求,解决了采用传统的Bayer图像恢复算法所产生的边界模糊以及局部图像有团块等问题.

This paper presents a Bayer image restoration system with the architecture of dual-processor pipelines： digital signal processor（DSP） and field programmable gate array（FPGA）, which can be applied to high-speed image acquisition and data processing. Firstly, the DM642 executed core processing algorithm and returned it to the FPGA for data preprocessing to recover the lost pixels. Then, an image in a certain format was packaged and transferred to the DM642 for the bilinear interpolation operation. The results show that the system can process a Bayer image of 752 × 480 pixels within15 ms, and can solve traditional problems such as edge blurring and lumps in images.