Method for achieving positioning by using wheel type odometer-IMU and monocular camera

Abstract

The invention discloses a method for achieving positioning by using a wheel type odometer IMU and a monocular camera. The method mainly comprises the following steps: (1) calibrating camera internal parameters, camera and wheel type odometer external parameters, and IMU and camera external parameters; (2) selecting an image key frame from an image shot by a camera, correcting distortion of the image, extracting feature points, performing system initialization, and finally selecting a current frame pose when initialization succeeds as a pose of a world coordinate system; (3) projecting the map point to the plane image according to the estimated value of the vehicle body pose; (4) predicting the pose of the vehicle body; (5) establishing a local map; (6) establishing an optimization problem about all poses and map points of the vehicle body according to the local map, and then obtaining the optimized poses and map points through a bundle adjustment method; and (7) detecting loopback by adopting a bag of words (BoW), and performing loopback correction once detecting that the same position is passed.

Description

technical field [0001] The invention belongs to the technical field of computer vision, and mainly studies a method for realizing positioning by using a wheel-type odometer-IMU and a monocular camera. Background technique [0002] With the development of SLAM technology, the positioning and navigation of ground vehicles began to realize positioning by laying some media on the ground (such as magnetic tracks, RFID, etc.), such as [Zhang Chenbeixi, Huang Zhiqiu, A review of the development of automatic guided vehicles (AGV)[J] Information Technology Application. 2010.], and now it is gradually changing to a trackless positioning method. Due to the relatively low cost of the camera and the images it captures can provide rich information, more and more attention has been paid to using visual SLAM (simultaneous localization and mapping) to achieve positioning and navigation, and many visual SLAM solutions have emerged, such as RGB_D-based cameras The SLAM scheme RGB_D-SLAM [Enge...

AI Technical Summary

Problems solved by technology

[0003] However, there are still some problems in the scheme of directly using visual inertia on ground vehicles. When the vehicle moves at a constant speed, the inertial measurement unit cannot detect the acceleration, which causes the scale uncertainty to increase at this time, which in turn leads to scale drift and positioning error. big

Images