一种CMOS相机色彩还原算法

Color reversion algorithm of CMOS camera

在靶场光学测量中为了兼顾目标跟踪测量和实况记录,提出一种使用CMOS相机进行色彩还原的方法。设计了基于该方法的硬件平台,并对软件算法的还原质量进行了定量分析。介绍了基于彩色滤波阵列的Bayer滤波算法原理,并建立了数学模型。介绍了双线性插值算法并分析了算法在图像边缘会造成拉链效应的不足,加入了边缘检测方法以克服该缺点。介绍了基于该算法的硬件平台和软件设计。实验结果表明,基于最高峰值比评价方法,还原效果可提高14%左右,时间性能满足实时性要求。改进的算法可以兼顾实时跟踪测量和实况记录的要求,客观判据和主观观感均表明还原效果较好。

In order to take advantages of tracking and real time record, a color reversion algorithm based on CMOS camera is proposed. The hardware platform is designed and reversion quality is analyzed quantitatively. First, the Bayer filter method based on color filter array is presented and mathematics module is designed. Then the bilinear algorithm is introduced and the edge detection method is added in order to reduce the zip effect when the edge of an image is crossed. The hardware and soft- ware is introduced. Experimental results indicate that the algorithm can offer a better PNSR which is 14% increased than the bilinear algorithm while the real time processing is achieved. The algorithm can give consideration to tracking and real time record, the objective calculation and subjective judgement indicate a better visual effect.