Visual simultaneous localization and mapping method based on depth convolution auto-encoder

Abstract

The invention discloses a visual simultaneous localization and mapping (Visual-SLAM) method based on a depth convolution auto-encoder. The method comprises the steps of 1, performing data preprocessing on training data; 2, establishing a multi-task learning network; 3, taking three adjacent frames of binocular images in the image sequence as network input; 4, constructing a loss function; 5, training, verifying and testing the multi-task network; 6, the trained shared encoder network is used for loopback detection; 7, constructing a new Visual-SLAM system front end through the six steps, constructing a rear end of the Visual-SLAM system through pose graph optimization or factor graph optimization, and building a complete system, and 8, verifying the positioning accuracy and robustness. A depth convolution auto-encoder is used, a semi-supervised multi-task learning method is used to construct the front end of an SLAM system, depth estimation, camera pose estimation, optical flow estimation and semantic segmentation are included, and a feature map of a network is used to construct image representation to perform loop detection.

Description

Technical field [0001] The invention belongs to the technical field of image processing; in particular, it relates to a method for simultaneous visual positioning and map construction based on a deep convolutional autoencoder. Background technique [0002] Simultaneous positioning and map construction technology means that a mobile robot equipped with a specific sensor uses sensors to recover the three-dimensional information of the scene during the movement process without a priori conditions of the environment, and the key technology to locate its own pose is to realize the robot path planning, Basic requirements for autonomous navigation and other complex tasks. [0003] A complete visual simultaneous positioning and mapping (Visual-SLAM) system can theoretically be divided into two parts: the front end and the back end, the relationship diagram is as follows figure 1 Shown. The front-end part mainly includes visual odometer, local map construction and loop detection. The visu...

AI Technical Summary

Problems solved by technology

The true value of the camera pose can be obtained by using GPS, IMU or multi-sensor fusion, but the absolute error of GPS is relatively large and cannot be used indoors or in outdoor environments with a lot of occlusion, even if multi-sensor fusion is used, it is not easy to obtain accurate pose ground truth

Images