Indoor Robot Simultaneous Localization And Mapping Method Research

Simultaneous Localization and Mapping(SLAM)technology is an important field in robot science,and it has been widely used in many fields such as manufacturing,agriculture,medical and health industry,service industry and national defense industry.Among them,the field of indoor robot is one of the hot spots.However,because of its high cost of calculation on the extraction and matching features,the algorithm of the traditional indoor robot RGB-D SLAM is not desirable for real-time performance.At the same time,due to the structural characteristics of the depth camera,the depth information at the edge can not be fully acquired,and most of the features are at the edge of the object.This situation makes it difficult for system to find enough characteristic points when calculating the robot position,which is likely to cause the loss of pose.In this paper,an improved ORB algorithm is proposed to combine significant feature filtering.In this paper,the ORB feature is used to reduce the computational complexity,and the scale space is constructed by the box filter,which makes the scale invariance of the ORB be fixed.In this paper,the ORB feature is used to reduce the computational complexity.Ascension.In order to solve the problem of error matching,this paper proposes a feature point screening method which combines the visual significance characteristics.In this paper,we use the efficient spatial graph calculation method based on color statistics to calculate the significant value of each key point.Through the feature point itself and the significant value of the surrounding points to compare the matching of the feature points,the gap is too large to remove the screen to improve the matching accuracy and speed..In this paper,we propose a method to calculate the robot pose in RGB-D SLAM by far point classification and beam balance method.By dividing the feature points according to the distance,the depth is obtained directly for the depth of the trusted point;for the depth of the untrusted point of the multi-frame method to estimate thedepth of information,so rich in the map of the key points.The beam correction method is used to reconstruct the spatial point,and the rotation and translation of the robot are optimized and solved by the error function.At the same time,the improved ORB feature and the explicit matching are used to construct the SLAM system,which makes the posture calculation The lack of a certain degree of robustness..This paper has been verified on the SLAM public data set of TUM.And the experimental results show that the proposed method can effectively make simultaneous location and mapping in indoor environment.For the dynamic fuzzy scene,the method of this paper can also get desirable results.