一种煤矿井下图像增强算法  

An image enhancement algorithm for coal mine underground

作　　者：程德强[1] 郑珍[1] 姜海龙[1] 

机构地区：[1]中国矿业大学信电学院,江苏徐州221008

出　　处：《工矿自动化》2015年第12期31-34,共4页Journal Of Mine Automation

基　　金：江苏省产学研前瞻性联合研究项目(SBY2013019-06);国家自然科学基金委员会与神华集团有限责任公司联合资助项目(U1261105)

摘　　要：针对煤矿井下监控图像对比度低、视觉效果差等问题,采用一种基于引导滤波的单尺度Retinex算法进行煤矿井下图像增强。该算法采用具有保边平滑功能的引导滤波进行照度估计,并与原图像进行卷积运算,获得反映自身属性的反射分量,然后针对经反对数变换的反射分量,采用改进的局部对比度增强方法提高图像暗区的对比度,最后得到增强图像。实验结果表明,该算法在对比度提升、细节增强、消噪方面取得了较好的效果。In view of low contrast ratio and view effect of monitoring images in coal mine underground,a single scale Retinex algorithm based on guided filtering was proposed for image enhancement of coal mine underground.The guided filtering,which can protect edge and make image smooth,is used for illumination estimation and convolution operation with original images,so as to get reflection component which reflects properties of the original images. For reflection component by anti-logarithm transformation,an improved local contrast enhancement method is used to improve contrast ratio of dark parts in the reflection image,so as to get enhanced image finally.The experimental results show that the algorithm has better effect in promoting contrast ratio,enhancing details and denoising.

关 键 词：煤矿井下 监控图像 图像增强 单尺度Retinex 引导滤波 

分 类 号：TD67[矿业工程—矿山机电]