一种红外焦平面阵列盲元检测与补偿方法  

A Method of Bad Pixel Detection and Compensation for IRFPA

作　　者：胡鹏博[1] 詹东军 陈华旺[1] 

机构地区：[1]华中光电技术研究所-武汉光电国家实验室,湖北武汉430223

出　　处：《光学与光电技术》2017年第3期69-73,共5页Optics & Optoelectronic Technology

基　　金：国家重大科学仪器设备开发专项(2013YQ290489)资助项目

摘　　要：针对红外焦平面阵列探测器成像存在盲元的问题,从工程应用角度出发,提出了一种易于在硬件上实现的盲元检测与补偿方法。对红外图像建立直方图模型,模拟高斯正态分布,利用"3σ判断准则"进行盲元检测;通过对图像合理分块,保证了盲元检测算法的快速性和鲁棒性。最后,结合红外图像边缘信息,提出了一种可靠的盲元补偿方法,可有效解决盲元在图像边缘补偿不稳定的问题。仿真结果表明,该方法可以有效地处理盲元,明显提高了红外成像的视觉效果。For the problem of bad pixel in IRFPA, a method of bad pixel detection and compensation based on hardware is described for engineering application. ＂The 3σ rule＂ is introduced to detect the bad pixel of the histogram that build on the infrared image. The histogram model is developed to simulate Gaussian normal distribution. The method of bad pixel detec- tion is fast and robustness, according to breaking the image into blocks. Finally, a method to compensate the bad pixel is proposed on the basis of the edge information of image. And the method effectively solves the problem that the traditional bad pixel compensation is poor in the edge of the image. Simulation result shows the method above can solve the bad pixel and improve the visual effect of infrared imaging.

关 键 词：红外焦平面阵列 盲元 直方图 SOBEL算子 边缘检测 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]