首页> 中文期刊> 《哈尔滨理工大学学报》 >基于固定图形噪声统计特性异常的盲元检测方法

基于固定图形噪声统计特性异常的盲元检测方法

         

摘要

The fixed pattern noise component of 3D noise model is used to detect blind-pixel of IRFPA as data source. It can be classified to dead pixel and over hot pixel by representable form of blind-pixel. As for dead pixel,the paper uses double-end dividing threshold value to detect dead pixel, which is determined by prior knowledge using quantity limited and low probability density characteristic of its hitstogram. But as for over hot pixel, it is regarded as abnormal pixel of Gaussian distribution noise, and then, detected by statistic distance threshold value that is determined by asymmetry coefficient of hitstogram removed dead pixel. The result of experiments shows that it is a practical method of blind-pixel detection.%针对盲元对于红外辐射的响应具有很强非线性的特点,在对焦平面采用线性非均匀性校正之前,必须采取有效的盲元检测方法对盲元予以剔除或补偿等问题,提出了以三维噪声固定图形噪声分量作为红外焦平面盲元检测的数据源.盲元的表现形式分为死元与过热像元.针对死元,将其数量限制作为先验知识,并结合其在数据直方图中的低概率密度特性,确定双端分割阈值进行检测.针对过热像元,将其作为正态分布噪声中的统计异常像素点,利用死元剔除后图像数据的直方图偏度阈值自适应确定统计距离门限,并进行检测.实测试验验证了方法的有效性.

著录项

相似文献

  • 中文文献
  • 外文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号