CAJ | 학술논문

在分析盲元特性的基础上，依据形态学开闭运算对尖峰信号的滤波性质，提出了一种新的红外焦平面阵列盲元检测算法。选取合适的结构元素，运用开闭运算的滤波作用对图像进行平滑处理，从平滑图像中减去其均值得到阵列噪声响应图像，选取合适的图像均值的倍数作为检测阈值。对阵列图像进行开闭运算并与平滑图像取差后得到对过亮及过暗像素增强的图像，每个像素与阈值进行比较判别出盲元。
재분석맹원특성적기출상，의거형태학개폐운산대첨봉신호적려파성질，제출료일충신적홍외초평면진렬맹원검측산법。선취합괄적결구원소，운용개폐운산적려파작용대도상진행평활처리，종평활도상중감거기균치득도진렬조성향응도상，선취합괄적도상균치적배수작위검측역치。대진렬도상진행개폐운산병여평활도상취차후득도대과량급과암상소증강적도상，매개상소여역치진행비교판별출맹원。
A new algorithm of blind pixel detection for infrared focal array is presented according to filtering characteristic of morphologic open and close operations for acuminate signal based on analysis of blind pixel character. Smoothing image is obtained by means of filter character of open and close operations with proper structure element. The noise response image of the array is gained by subtracting the mean of all pixels from each pixel of the smoothing image. The mean is designated as threshold of defective pixels detection. Gray scale open and close operators are applied to the array respectively, then the outcome image subtracts the smoothing image to produce the image with enhanced over hot pixels and dead pixels. Each pixel is compared with the threshold, and blind pixels are differentiated.