一种基于Retinex的矿井非均匀照度图像增强算法

煤矿井下的安全高效生产离不开清晰流畅的监控图像,由于矿井下粉尘多、照度低或点光源照射而形成特殊的非均匀照度环境,使得矿井监控图像整体偏暗,对比度低,带有很强的背景噪声,视觉效果差。针对这一问题,基于Retinex算法,提出了一种在高斯滤波后用迭代矢量法继续分离亮度图像的方法。实验结果表明该方法能有效的增加图像对比度,降低背景噪声,改善图像的整体视觉效果。

Algorithm of Mine non-uniform illumination image enhancement based on Retinex

Clear and smooth surveillance images are necessary for knowinging the coal mine safe and efficient production in the ground control room, because of the more dust in mine, low illumination or point light source to form a special non-uniform illumination environment, the mine monitoring images are all dark,with low contrast and strong background noise. The visual effect is poor. In response to this problem, the algorithm, that continues use iterative vector method after Gaussian to separates the luminance image is proposed. Experimental results show that this method can effectively increase the image contrast and reduce background noise and improve the overall visual effect of the image.