基于统计特性的路面图像光照不均匀校正算法

针对传统的灰度校正方法不能有效地校正全局光照不均匀的路面图像,提出基于统计特性的路面图像光照不均匀校正算法。算法根据路面图像自身的特点,以及正态分布特性,利用图像的均值和方差估计每个子块图像的背景灰度值,然后利用双线性插值的方法构造背景灰度矩阵,最后原图像减去背景灰度即可得到校正后的图像。仿真实验证实,该算法能有效地减少图像的光照不均匀,达到校正图像的目的。

the Algorithm of Correcting Uneven Illumination Pavement Image Based on Statistical Properties

In view of the traditional gray correction method little effect on correcting global illumination uneven pavement image, it proposed a method of correcting uneven illumination pavement image based on statistical properties. According to the characteristics of pavement image as well as their normal distribution, The algorithm estimats the background gray each subblock image using image mean and variance, and then creates background gray matrix using the bilinear interpolation method, finally gets the corrected image using the original image and the background image subtraction. Simulation results show that the algorithm can effectively reduce image non-uniform illumination and correct the image.