改进Retinex无人机图像去雾算法  

Modified Image Defogging Algorithm for Retinex UAV

作　　者：许哲[1] 吴琨 黄鹤[2] 黄莺[3] 郭璐 XU Zhe;WU Kun;HUANG He;HUANG Ying;GUO Lu(No.20 Institute of CETC,Xi’an Shaanxi 710068,China;Chang’an University,Xi’an Shaanxi 710064,China;Air Force Engineering University,Xi’an Shaanxi 710038,China;Xi’an Aisheng Technology Group Co.Ltd.,Xi’an Shaanxi 710075,China)

机构地区：[1]中国电子科技集团公司第二十研究所,陕西西安710068 [2]长安大学,陕西西安710064 [3]空军工程大学,陕西西安710038 [4]西安爱生技术集团公司,陕西西安710075

出　　处：《通信技术》2020年第12期2942-2946,共5页Communications Technology

基　　金：陕西省自然科学基础研究计划面上项目(No.2019JM-610);陕西省创新人才推进计划——青年科技新星项目(No.2019KJXX-028);长安大学中央高校教育教学改革专项经费资助项目(300103103211);长安大学2019年创新创业教育研究与培训中心建设项目(2019YP01)。

摘　　要：无人机航拍技术以其诸多优势,被广泛应用于图像处理领域。但是,无人机实际飞行过程中易受到雾霾的影响,致使获取的图像明显降质,而传统去雾算法处理效果不明显,在亮度较大的地方图像处理效果不佳,光晕现象突出。针对该问题,提出一种基于Retinex改进的无人机图像去雾算法。在Retinex算法基础上进行多尺度方法改进,通过多尺度视网膜增强算法保持一致性色感,利用直接线性量化优化输出图像品质,降低图像失真度,并极大地减少了光晕现象。实验结果表明,改进算法有效解决了存在的不足,鲁棒性较好,去雾效果明显。UAV aerial photography technology has been widely used in the field of image processing for its many advantages.However,it is easy to be affected by fog and haze in the actual flight process,resulting in obvious degradation of the acquired image.The processing effect of traditional defogging algorithm is not obvious,and the image processing effect is poor in places with large brightness,resulting in serious halo phenomenon.To solve this problem,a modified image defogging algorithm for UAV based on Retinex is proposed.Based on the Retinex algorithm,the multi-scale method is modified to maintain the consistent color sense through the multi-scale retinal enhancement algorithm.The direct linear quantization is used to optimize the output image quality,reduce the image distortion degree and the halo phenomenon.The experimental results indicate that this modified algorithm can overcome the above shortcomings effectively,has good robustness and obvious effect of fog removal,and thus is effective and feasible.

关 键 词：无人机 去雾 RETINEX 直接线性量化 

分 类 号：TN911.73[电子电信—通信与信息系统]