Bayer彩色自动白平衡算法设计及其FPGA实现  

Design and FPGA implementation of Bayer color automatic white balance algorithm

作　　者：王达 赵冬青 储成群 WANG Da;ZHAO Dongqing;CHU Chengqun(Key Laboratory of Instrumentation Science&Dynamic Measurement,Ministry of Education,North University of China,Taiyuan 030051,China)

机构地区：[1]中北大学仪器与电子学院,太原030051

出　　处：《激光杂志》2021年第10期74-77,共4页Laser Journal

摘　　要：针对机器视觉对图像的高品质和实时性需求,提出一种基于FPGA的改进自动白平衡算法。该算法根据Bayer阵列的结构特点,采用双线性插值算法进行彩色图像复原,通过统计一帧图像中各通道像素均值和最大值求解校正系数,进而对原始图像进行校正。同时将该算法移植到FPGA中,利用FPGA实现高效并行结构特点,实现Bayer图像的实时白平衡处理。测试结果表明,在对分辨率为1 292×964的图像进行处理的情况下,达到了31帧/s的实时处理速度。对高曝光、低曝光等情况下的偏色图像进行校正分析,经过处理后图片的色差均值远小于原图的色差均值,与灰度世界法等经典算法相比,该算法所得到的色差均值也足够小,在主观和客观评价上都取得了较好的效果,能够有效提高图像品质,具有广阔的应用场景。Aiming at the high quality and real-time requirements of machine vision,an improved automatic white balance algorithm based on FPGA is proposed. According to the structure characteristics of Bayer array,the algorithm uses bilinear interpolation algorithm for color image restoration,calculates the correction coefficient by counting the average value and maximum value of each channel pixel in a frame image,and then corrects the original image. At the same time,the algorithm is transplanted to FPGA to realize the real-time white balance processing of Bayer image. The test results show that the real-time processing speed of 90 frames/s is achieved in the case of 1 280 × 720 resolution image processing. After correcting and analyzing the color biased image in high exposure and low exposure,the average color difference of the processed image is far less than that of the original image. Compared with the gray world method and other classical algorithms,the average color difference obtained by the algorithm is small enough,which achieves good results in subjective and objective evaluation. It can effectively improve the image quality and has broad application prospects.

关 键 词：FPGA 实时图像处理 自动白平衡算法 颜色插值 

分 类 号：TP394.1[自动化与计算机技术—计算机应用技术] TH691.9[自动化与计算机技术—计算机科学与技术]