摘要
针对视觉同步定位与地图构建算法在低光照和高光照场景下会定位失败和跟踪丢失的问题,提出一种自适应复杂光照的SLAM算法;提出一种全局自适应图像增强算法,利用改进的自适应全局色调映射和自适应gamma变换调节亮度,再使用CLAHE方法调节对比度;提出自适应阈值算法计算FAST角点检测阈值。实验结果表明,与原算法相比,在相应数据集的4个序列上,绝对轨迹误差的均方根误差平均降低了49.85%。
To address the issues of localization failure and tracking loss in visual simultaneous localization and mapping(SLAM)algorithm under low and high illumination conditions,an adaptive SLAM algorithm for complex lighting is proposed.Firstly,a globally adaptive image enhancement algorithm is proposed to adjust brightness by using improved adaptive global tone mapping and adaptive gamma correction,and adjust the contrast by using the CLAHE method.Secondly,an adaptive threshold algorithm is proposed to calculate the FAST corner point detection threshold.The experimental results show that compared with the original algorithm,the root mean square error of absolute trajectory error on the four sequences of the corresponding dataset is reduced by an average of 49.85%.
作者
吕艳辉
钟强
LYU Yanhui;ZHONG Qiang(College of Computer Science and Engineering,Shenyang Ligong University,Shenyang 110159,China)
出处
《火力与指挥控制》
北大核心
2025年第8期114-122,共9页
Fire Control & Command Control
基金
辽宁省教育厅基本科研基金资助项目(JYTMS20230192)。
