【摘要】 微光电视是低照度条件下的重要观测手段 ,但微光电视图像信噪比低 ,观察效果差 ,必须进行降噪处理 根据噪声小波系数随着小波分解尺度加大而减小的特性 ,文中算法利用微光图像小波分解层间相关性 ,可以良好地消除噪声 ,同时克服了传统降噪方法在滤除噪声的同时会造成图像细节损失的缺点 从所介绍的微光图像实时处理系统实现结果看 ,降噪效果明显 本文还给出了算法工程化实现的定标和阈值确定方法更多还原

【Abstract】 Low Light Level (3L) TV is an important observation measure in the condition of low illumination. But the low SNR of it severely deteriorates the observation quality. An algorithm using correlation of wavelet coefficients based on the fact that amplitude of wavelet coefficients of noise decreases strongly when the scale increases is presented.The results of the experiments to apply the above mentioned algorithm to a real time 3L image processing system show satisfying effect. The noise of the resulting image has been suppressed whereas the detail has been less modified. Scaling and threshold choosing algorithms are also put forward.更多还原