摘要
针对热像仪普遍存在的暗角噪声进行了分析,提出了一种基于参考辐射源进行暗角噪声补偿的方法。探讨了传统非均匀性校正方法对暗角噪声进行补偿的优劣性,并提出了一种改进的神经网络校正算法。基于Retinex理论提出了一种基于单帧图像快速去除暗角噪声的方法,实验结果表明,该算法可以有效去除暗角噪声。
A dark-corner compensation method based on reference source is proposed in this paper, following the analysis of the multi-view thermal imager dark-corner noise. The superiority-inferiority of NUC algorithm is discussed based on multi-frame images to solve the problem of the dark-corner noise, and improve the NUC algorithm based on neural network. A fast algorithm of removing dark-corner noise based on single frame image is also introduced in this paper. The experimental results show that the proposed algorithm can remove dark-corner noise effectively.
作者
粟宇路
杨波
夏菲
苏兰
苏俊波
刘传明
陈大乾
SU Yulu YANG Bo XIA Fei SU Lan SU Junbo LIU Chuanming CHEN Daqian(Kunming Institute of Physics, Kunming 650223, China State Grid Liaoyang Electric Power Supply Company Information ＆ Telecommunication Branch, Liaoning 111000, China)
出处
《红外技术》
CSCD
北大核心
2017年第3期226-231,236,共7页
Infrared Technology
