Research On Simultaneous Localization And Mapping Algorithm Of Combination Optimization Based On RGB-D Camera

In the unknown environment,the robot needs to accurately know its position in the location environment when moving,and establish a simple environment map.Simultaneous localization and mapping(SLAM)is a very important technology in robotic or autonomous driving applications.Its main goal is to complete positioning while autonomously completing positioning without external intervention.In recent years,the radar-based laser SLAM technology has achieved great success in many fields,but due to its high cost and few features,its application is still limited;The camera-based visual SLAM technology has the characteristics of low cost,rich features,and accurate reflection of the real environment,and the research has also attracted more and more attention.In the visual SLAM,there are many challenges in traditional point feature-based approaches,such as insufficient accuracy,motion jitter and tracking failure,which reduce the performance of the algorithms.In response to these issues,the thesis makes some research and improvement:First of all,this thesis proposes a novel algorithm called visual SLAM to handle these problems.we present the bilinear interpolation method to get the gray values of the feature points,the estimated pose of the current frame can be obtained by minimizing the photometric error between two frames.In order to further reduce the pose error,the pose of the current frame is optimized by minimizing the reprojection error,so as to improve the localization accuracy.We also come up with a new keyframe selection mechanism to improve the mapping accuracy,which measures the motion amplitude from the previous key frame to the current frame according to the pose,and judges whether the current frame is added to the keyframe sequence to rationalize the selection of keyframes.Secondly,it is easy to fall into the local optimal value when minimizing the photometric error of two images.In this thesis,when optimizing the photometric error,the image pyramid method is used to optimize the position and posture from the top of the pyramid to the bottom layer by layer.At the same time,the pose optimization results of the upper layer are taken as the initial values of the next layer,which is a process of gradually reducing the error,And when the camera displacement is large,it can still track the pixels.This thesis thoroughly evaluates the system on the TUM RGB-D benchmark and compares it with ORB-SLAM2,our method has better accuracy and robustness.