基于HSI空间融合的矿井图像增强算法  

Mine image enhancement algorithm based on HSI space and fusion

作　　者：朱礼义 李巧月 李国超 郭小兵 韩习习 祝汉城 ZHU Li-yi;LI Qiao-yue;LI Guo-chao;GUO Xiao-bing;HAN Xi-xi;ZHU Han-cheng(School of Information and Control Engineering,China University of Mining and Technology,Xuzhou 221008,China;Peng Zhuang Coal Mine,Linyi Mining Group Heze Coal Power Limited Company,Heze 274700,China)

机构地区：[1]中国矿业大学信息与控制工程学院,江苏徐州221008 [2]临沂矿业集团菏泽煤电有限公司彭庄煤矿,山东菏泽274700

出　　处：《计算机工程与设计》2019年第10期2926-2930,3008,共6页Computer Engineering and Design

基　　金：“十三五”国家重点研发基金项目(2016YFC0801808);江苏省科技成果转化基金项目(BA2017009)

摘　　要：针对煤矿井下图像光照不均匀、照度低、视觉效果差等特点,提出一种基于HSI空间变换和融合的井下彩色图像增强算法。将源图像进行HSI变换得到亮度分量I,对分量I进行同态滤波处理得到分量I';分别对分量I和I'进行Curvelet变换分解,对两者高频系数采用拉普拉斯能量和(SML)取大的方法融合,对两者低频系数采用局部能量方差积法进行融合;通过Curvelet反变换还原出亮度分量IF,采用局部对比度增强方法对IF处理,提高图像的对比度;将处理后的分量IF和分量H、S合成增强图像。实验结果表明,该算法在消除光照不均匀、对比度增强、噪声消除等方面效果较好。In view of poor visual,uneven and low illumination of mine images,a downhole color image enhancement algorithm based on HSI transform and fusion is proposed.The source image was transformed by HSI transform to get the brightness component I,and the component I was processed by homomorphic filtering to get the component I'.The component I and I' were decomposed by Curvelet transform respectively,and the high-frequency coefficients were fused using sum-modified-Laplacian(SML)maximum method,and the low frequency coefficients were fused using region energy and variance product method.The luminance component IF was restored by Curvelet inverse transform,and the region contrast enhancement method was used to improve the contrast of the image.The enhanced image was obtained by HSI inversion through the component IF,Hand S.Experimental results show that the proposed algorithm has better performances in uneven illumination elimination,contrast enhancement and noise cancellation.

关 键 词：矿井图像 光照不均匀 图像增强 CURVELET变换 同态滤波 

分 类 号：TD391[矿业工程—矿井建设]