Binocular image-based visual map generation method for mobile robot

Abstract

The invention discloses a binocular image-based visual map generation method for a mobile robot. The method includes: performing visual matching and loop detection on related images according to visual features of a binocular sequence image, and in order to better recover a scene scale, fully utilizing binocular baseline constraints to improve a parameterized expression mode of point cloud, and designing a mapping initialization algorithm, an incremental algorithm and a corresponding global smooth adjustment process designed for a binocular camera. Compared with an existing binocular mapping method, the method is higher in speed and lower in complexity; compared with an SLAM-based mapping algorithm, map drifting is smaller, scene robustness is better achieved, the adaptive capacity of a complex environment is enhanced, the scene composition requirement of an existing robot is met, meanwhile, high expandability is achieved, the whole mapping link step can be accelerated in combination with the output result of the SLAM, and meanwhile, the robustness can be further enhanced.

Description

technical field [0001] The invention belongs to the technical field of visual maps, and in particular relates to a method for generating a visual map of a mobile robot based on a binocular image. Background technique [0002] Recovering the structure of a scene from a series of pictures to build a map for localization is one of the key problems in robot navigation. Recovering the scene structure from motion is to use the observation information generated by the motion to the sensor, and gradually use the information of the multi-view geometry to calculate the position of the sensor itself and restore the coordinates of the map point. The challenges are as follows: [0003] (1) The traditional visual SLAM (Simultaneous Localization and Mapping) algorithm is often highly dependent on texture in terms of accuracy and robustness, but also has problems such as poor accuracy, insufficient loop closure, and cumulative error in attitude. In some extreme cases (such as machine coll...

AI Technical Summary

Problems solved by technology

However, this mapping method still has a large amount of attitude drift. At the same time, due to multiple smoothing calculations in the incremental process, not only the calculation results are not good, but the process is often time-consuming.

Images