Vision-based offline map establishing and positioning method

Abstract

The invention relates to a map establishing method. The map is applied to a vision positioning technology of a mobile robot. The method at least comprises the following steps: A, acquiring an RGB image and a depth image in an application occasion by a 3D camera acquisition robot and calculating parameters of three-dimensional point cloud; B, extracting the ORB feature of the acquired image and a descriptor, splicing the image of the three-dimensional point cloud of the surrounding environment and generating key frames of the RGB image and the depth image; C, generating a word bag according tothe extracted ORB feature and the descriptor and clustering to generate an ORB feature tree; D, analyzing the space direction in the three-dimensional point cloud, acquiring projection parameters of the three-dimensional point cloud space, projecting the point cloud vertical to the ground to the ground according to requirements to form a 2D map; and E, acquiring a positioning map required by vision positioning of a mobile robot and key frames of the map. In addition, the invention also provides a robot positioning method using the map establishing method. By the method, the problems that positioning is in failure and the position of the robot cannot be known in the prior art are solved.

Description

technical field [0001] The invention relates to robot positioning technology, in particular to a vision-based off-line map construction and positioning method. Background technique [0002] At present, with the development of mobile robot technology, more and more mobile robot products have entered the consumer market and commercial market. In the market field of commercial applications, mobile robots are mainly used for guidance, shopping guide, explanation, introduction, welcome, and item delivery. As a mobile platform, mobile robots have mobility as their basic function, and the positioning function of mobile robots is one of the key technologies for mobile robots to achieve autonomy. Mobile robots can only understand their own position in the current environment. , to decide where to go next? And how to get there? There are many positioning methods for mobile robots, such as: WIFI positioning, RFID positioning, UWB positioning, visual positioning, iBeacon positioning,...

AI Technical Summary

Problems solved by technology

[0005] The object of the present invention is to provide a vision-based off-line map construction and positioning method. Based on the map construction method proposed by the present invention, the positioning of the mobile robot is based on vision, IMU (Inertial measurement unit, inertial measurement unit) and odometer. The combined positioning method solves the technical problem that the robot cannot know its own position and cannot navigate after the failure of the robot's visual positioning, and improves the accuracy of visual positioning

Images