针对传统的双线性Bayer彩色图像恢复算法恢复效果不理想，边界部分不明显，局部图像有团块等问题，提出了一种基于硬件实现的Bayer图像快速插值算法。采用梯度算法对Bayer格式图像绿色通道进行恢复，根据像素点所属的颜色组对蓝色通道进行恢复。实验结果表明，本文算法比双线性法有更好的峰值信噪比(PSNR)值，RGB 3个通道的PSNR值均比双线性法高5 dB以上，而且算法消耗时间比双线性法少，恢复的图像视觉效果更好。实验处理一幅512×512的全彩图像仅需要9.3 ms，完全可以满足实时性的要求，因此,本文算法在对实时性要求高的场合有很好的应用前景。

An improved fast Bayer interpolation algorithm based on the Field Programming Gate Array(FPGA) was presented to solve the problems of a traditional bilinear Bayer color image restoration on edge blurring and lumps in images. With the algorithm， a gradient algorithm was used to restore the green channel and the color groups corresponding to pixels to restore the blue channel for a image. Experimental results indicate that the algorithm can offer a better visual effect of the images, and the Signal to Noise Ratios(SNRs) of RGB channels obtained from the proposed algorithm are all higher 5 dB than those of obtained from the bilinear algorithm. Moreover, the time-consuming of the algorithm is less than that of the bilinear algorithm, and a 512× 512 full color image can be processed in 9.3 ms. These results can meet the requirements of real time image processing and have higher application values.