弱光环境下视觉SLAM前端优化方法

Optimization Method of Front End of Visual SLAM in Weak Light Environment

为了解决机器人在弱光环境下可能出现的自身定位不准确及建图不完整的情况,采用了一种基于弱光环境下图像增强的方法对相机采集到的图像进行预处理,提高图像的亮度和清晰度,对增强后的图像用ORB算法提取和匹配特征点,使用Hamming距离筛选法进行特征点的误匹配处理。实验结果表明:本优化算法有效地增加了弱光环境下的特征点提取数量,并且经本算法优化后的三维点云图质量更高。

In order to solve the problem of inaccurate positioning and incomplete mapping of robot in weak light environment,we used an image enhancement method based on weak light environment to preprocess the images collected by the camera and improved the brightness and clarity of the images.The images feature points were extracted and matched by using the ORB algorithm,and the Hamming distance screening method was used for processing mis-matched feature points.The experimental results show that the proposed algorithm effective-ly increase the number of feature points extraction in weak light environment,and the quality of 3Dpoint cloud image optimized by this algorithm is higher.