基于图像增强的ORB-SLAM2算法水下应用研究  

Research on underwater application of ORB-SLAM2 algorithm based on image enhancement

作　　者：姚三坤 刘明 陈世强 YAO San-kun;LIU Ming;CHEN Shi-qiang(College of Electrical and Information Engineering,Yunnan Minzu University,Kunming 650504,China)

机构地区：[1]云南民族大学电气信息工程学院,云南昆明650504

出　　处：《云南民族大学学报（自然科学版）》2023年第6期785-792,共8页Journal of Yunnan Minzu University:Natural Sciences Edition

基　　金：国家自然科学基金(52061042)。

摘　　要：为了解决水下视觉SLAM算法由于图像对比度低、模糊等导致的特征提取鲁棒性差的问题,采用自适应直方图均衡化算法作为图像增强算法作用在ORB-SLAM2的前端,提高了水下图像对比度和清晰度的同时保留了原图的细节区域,使得质量不高的水下图像也能较好的提取特征.通过在数据集上的实验表明自适应直方图均衡化对水下图像的特征提取及关键帧跟踪效果明显改善,图像增强的ORB-SLAM2算法在水下环境中能较好的进行特征提取及跟踪,定位不易丢失,SLAM算法准确性和鲁棒性明显提高.In order to solve the problem that underwater SLAM algorithm has poor robustness of feature extraction due to low image contrast and blur,the adaptive histogram equalization algorithm is used as the image enhancement algorithm in the front end of ORB-SLAM2,which improves the contrast and clarity of underwater image while retaining the details of the original image.So that the underwater image with low quality can be better feature extraction.Experiments on data sets show that the effect of adaptive histogram equalization on feature extraction and key frame tracking of underwater images is significantly improved.The IMAGE enhanced ORB-SLAM2 algorithm can perform feature extraction and tracking better in underwater environment,and localization is not easy to lose.The accuracy and robustness of SLAM algorithm are significantly improved.

关 键 词：水下视觉SLAM 图像增强 特征提取 自适应直方图均衡化 ORB-SLAM2 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]