A binocular vision indoor positioning and mapping method and device

Abstract

The invention discloses a binocular vision indoor positioning and mapping method and device, comprising: collecting left and right images in real time, and calculating the initial pose of the camera; collecting angular velocity information and acceleration information in real time, and pre-integrating to obtain the state of the inertial measurement unit Quantity; construct a sliding window containing several image frames, constrain the visual error term between the image frames and the error term of the inertial measurement unit measurement value, conduct nonlinear optimization on the initial pose of the camera, and obtain the optimized Camera pose and inertial measurement unit measurements; build a bag-of-words model for loopback detection, and correct the optimized camera pose; extract left and right image features and convert them into words and match them with the bag-of-words of the offline map. If successful, the solution will be optimized Otherwise, re-collect the left and right images and perform bag-of-words matching. The invention can realize positioning and mapping in an unknown environment, as well as a positioning function in a scene with a built map, and has better precision and robustness.

Description

technical field [0001] The invention relates to a binocular vision indoor positioning and mapping method and device, belonging to the technical field of positioning and navigation. Background technique [0002] At present, indoor robots are increasingly used in large shopping malls, warehouses and households, such as shopping guide robots in shopping malls, intelligent storage robots, and home sweepers. In these application fields, the robot needs to complete autonomous navigation, and to complete the autonomous navigation, the robot first needs to have the function of indoor positioning, that is to say, the robot needs to know its current location information in the scene and the location information of the destination. [0003] At present, there is no mature indoor high-precision positioning solution. The GPS (Global Positioning System, Global Positioning System) solution has a large indoor positioning error; the solution relies on pasting a QR code label in the scene to ...

AI Technical Summary

Problems solved by technology

[0006] The technical problem to be solved by the present invention is to solve the problems of high scene texture requirements, poor relocation ability and poor navigation effect in the existing indoor navigation mode, and provide a binocular vision indoor positioning and mapping method and device, through The indoor positioning method based on the fusion of visual and inertial navigation units can perform positioning and map building in complex scenes

Images