基于场景特殊区域的非均匀校正去鬼影算法

Nonuniformity correction deghosting algorithm based on special regions of scene

为了校正红外成像系统的非均匀性噪声,本文提出了一种基于场景特殊区域的非均匀性校正去鬼影算法。首先,采用三边滤波器对图像进行滤波,然后检测图像两种特殊的区域并对其校正系数进行调整。对检测到的高亮度区域,减小该区域相应的校正系数,使得该区域对校正系数的贡献减小;对检测到的灰度突变的边缘区域,增大该区域相应的校正系数,使得该区域对校正系数的贡献增大。通过这些步骤,可以有效避免校正系数的错误更新,可以有效减少鬼影现象。最后通过仿真实验证明,本文算法在去除场景特殊区域所造成的"鬼影效应"的效果优于空域低通时域高通算法和基于双边滤波器的时域高通滤波算法。因此,本文算法是有效的和实用的。

In order to correct the nonuniformity noise of the IR imaging system, a nonuniformity correction deghostingalgorithm based on special regions of scene was proposed. Firstly, the trilateral filter was used to filter the image, andthen the two kinds of special areas were detected, and their correction coefficients were adjusted. For the detectedhigh-brightness regions, correction coefficient of the region was reduced to make the region＇s contribution to the cor-rection coefficient decrease;For the edge region of the detected gray mutation, the correction coefficient of the regionwas enlarged to make the region＇s contribution to the correction factor increase. Through these steps, it can effectivelyavoid the correction coefficient to update falsely ,and it can reduce the ghost effect effectively. Finally ,the simulationexperiments show that this algorithm is better than space low-pass temporal high-pass algorithm and temporal high-pass filtering algorithm based on bilateral filter in removing the ghost effect caused by the special region. Therefore, thealgorithm is effective and practical.