摘要
在研究传统红外面目标滤除椒盐噪声算法的基础上,提出了多窗口自适应的面目标滤波算法。针对椒盐噪声在红外图像上是区域灰度值极值的这一特征,根据噪声密度的大小调整窗口区域的范围,找出窗口区域内灰度的最大和最小值像素点。在信噪比较高的情况下,利用窗口区域周围像素灰度值对噪声点重新赋值;在信噪比较低的情况下,选取阈值去除误检点,减少误差。仿真结果表明,该算法在高信噪比的情况下,平均绝对误差较传统算法降低了约1%～2%;在低信噪比的情况下,通过阈值处理,峰值信噪比较处理前提高约1%～4%。
On the basis of traditional salt and pepper noise filtering algorithm about infrared area target ,multi-window adaptive filter algorithm was presented. Because of salt and pepper noise is gray extreme value in infrared image, according to the noise density adjusted the scope of the region of window,selected the maximum and minimum value in the window area. In the ease of high SNR,using pixel gray value around the window region re-assignmened the noise point; in the case of low SNR,selecting proper threshold to remove false discreet and reduce error. Simulation results show that in high SNR conditions,the average absolute error than the conventional methos decreases by about 1%-2%; in the low SNR case, through threshold, peak signal to noise ratio than that of the former increases by about 1%-4%.
出处
《电子设计工程》
2010年第8期97-100,共4页
Electronic Design Engineering
关键词
红外图像
椒盐噪声
阈值
信噪比
infrared image
salt and pepper noise
threshold value
SNR
