摘要
首先分析了传统的非均匀性校正方法的缺点,指出自适应校正红外焦平面器件非均匀性的必要性。根据焦平面器件的非均匀性噪声特性和算法研究的需要,介绍了非均匀性失真图像的产生方法。在上述工作的基础上,研究了基于神经网络的自适应非均匀性校正算法,探讨了最近4邻域像素平均、最近4邻域像素灰度加权和8邻域像素灰度加权等三种情况。实验结果表明。
In this paper, the disadvantages of the traditional nonuniformity correction methods for infrared focal plane arrays are analyzed firstly. It is pointed out that adaptive nonuniformity correction for infrared focal plane arrays is necessary. Considered the nonuniformity noise characteristics of FPAs and in order to research algorithm, a simulation method of nonuniformity distortion image is introduced. Based on above works, the algorithms of nonuniformity correction for infrared FPAs using neural networks are researched. Three methods, which are the average of four nearest neighbors, the intensity\|weighted average of the nearest neighbor pixels algorithm and the intensity\|weighted average of eight\|neighbor pixels algorithm, are developed. Experiments show that the intensity\|weighted average of the eight\|neighbor pixels algorithm is effective.
出处
《红外与激光工程》
EI
CSCD
2000年第1期65-68,共4页
Infrared and Laser Engineering
基金
国防科工委基金
