针对仿生机器鱼水下作业时面临的水下图像质量偏低、水下自主定位难的问题,提出一种颜色均衡与G-B通道先验融合的水下图像增强式算法.将该算法和视觉同时定位与地图构建(SLAM)方法结合,实现了水下图像增强式的视觉三维重建.在不同水域环境下进行了水下图像处理实验、水下环境视觉三维重建实验和运动轨迹跟踪实验,结果表明该方法有效提高了水下图像综合质量,特征匹配效率提高了 16.03％,真实轨迹与估计轨迹的误差平均约为7.99 mm.
SLAM-based Underwater Image Enhanced Visual 3D Reconstruction Method
Regarding with problems of low quality of underwater images and difficulty of under-water autonomous localization faced by bionic robotic fish in underwater operations,an enhanced un-derwater image algorithm with color equalization and a priori fusion of G-B channels was proposed.The algorithm was combined with visual SLAM construction methods to enhance visual 3D recon-struction of underwater images.Underwater image processing experiments,underwater environment visual 3D reconstruction experiments,and motion trajectory tracking experiments were carried out in different water environments.The results show that the method effectively improves the comprehen-sive quality of underwater images.The feature matching efficiency is improved by 16.03％,and the er-ror between the real trajectory and the estimated trajectory is about 7.99 mm on average.
underwater image enhancement3D reconstructiontrajectory trackingvisual simul-taneous localization and mapping(SLAM)
万方数据
维普
