基于暗原色先验的煤矿井下退化图像复原算法  

Restoration Algorithms of Degradation Image in Underground Mine Based on Dark Channel Prior

作　　者：刘晓文[1] 仲亚丽[1] 袁莎莎[1] 黄河[1] 

机构地区：[1]中国矿业大学信息与电气工程学院,江苏徐州221116

出　　处：《煤炭科学技术》2012年第6期77-80,共4页Coal Science and Technology

基　　金：国家自然科学基金重点资助项目(60972059)

摘　　要：针对煤矿井下因水雾和煤尘散射作用引起的图像退化问题,结合煤矿井下无线多媒体节点采集的图像特点,提出一种正则化拉普拉斯矩阵的暗原色先验去雾尘模型。根据暗原色先验理论对来自煤矿井下无雾图像数据库进行统计,建立了煤矿井下图像成像的物理模型,利用该模型估算介质传播函数和井下光线照度,再由去雾尘模型复原得到清晰化的图像。试验结果表明,该算法有效恢复了场景的对比度,明显提高了图像的视见度。According to the image degradation problem caused by the water fog and coal dust scattering in the underground mine. In com- bination with the image features collected from the wireless multi-media nodes in the underground coal mine a regularization Laplacian Matrix of the dark channel prior fog and dust removing model was provided. According to the dark channel prior theory, a statistics was conducted on the no fog image data bank from the underground mine and a physical model of the underground mine imaging was estab- lished. The model was applied to estimate the medium transmission function and light illumination and the vivid images obtained would be restored with the fog and dust removing model. The experiment results showed that the algorithm could effectively restore the contrast ratio of the field and could obviously improve the visibility of the image.

关 键 词：暗原色先验 图像去雾尘 成像模型 介质传播函数 视见度 

分 类 号：TD76[矿业工程—矿井通风与安全]