基于方差补偿矩匹配的红外图像非均匀性校正方法  

Non-Uniformity Correction Method Based on Standard Deviation Value Compensation after Moment Matching

作　　者：祝善友[1] 张桂欣[1] 巩彩兰[2] 祝令亚[3] 赫华颖 胡勇[2] 

机构地区：[1]南京信息工程大学遥感学院,江苏南京210044 [2]中国科学院上海技术物理研究所,上海200083 [3]中国资源卫星应用中心,北京100094

出　　处：《光学学报》2013年第12期87-92,共6页Acta Optica Sinica

基　　金：民用航天"十二五"预先研究项目(D040103);国家自然科学基金(41001289);江苏高校优势学科建设工程

摘　　要：红外焦平面阵列的非均匀性校正是其定量应用之前必须解决的关键难题之一。将矩匹配方法引入到红外图像非均匀性校正中,针对其改变图像中地物光谱分布,并可能产生新条带噪声的缺陷,通过比较标准矩匹配前后每个探测元的输出与其均值差异,提出了一种方差补偿的矩匹配改进方法。实验图像校正结果的定性与定量分析表明,方差补偿的矩匹配方法效果理想,在校正图像非均匀性的同时,更大程度上保留了不同地物本身光谱特征的差异。Non-uniformity of infrared focal plane arrays is one of the key problems that must be solved before its quantitative applications. The moment matching method is adopted in correction of infrared non-uniformity. Due to the disadvantages of the method, such as changing of the image spectral variance and possibly producing new stripe noise, by comparing the output and the mean value difference for each detector between before and after the standard moment matching correction, an improved moment matching method based on the standard deviation value compensation is proposed. The quantitative and the qualitative analysis results show that the proposed method improves the non-uniformity correction precision, while the spectral variance of different objects in original image is better retained.

关 键 词：遥感 红外焦平面阵列 非均匀性校正 方差补偿 矩匹配 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]