A binocular vision environment detection system and method based on tactile perception

Abstract

The present invention relates to a binocular visual environment detection system and method based on tactile perception, the system at least includes a robot body for archaeological excavation, the robot body includes a camera device, a tactile sensor and a processor, and the processor is controlled by The configuration is: based on at least two visual images with parallax values ​​collected by the camera device, the statistical characteristic loss regions adjacent to each other are divided to form a curved path for the robot body to walk, and the curve paths related to the curved path are determined by the The tactile information collected by the tactile sensor is associated with the prior information to generate a set of statistical feature points on the surface of the object, so as to compensate the features of the lost area of ​​the statistical properties to extract image edge feature points.

Description

technical field [0001] The invention belongs to the field of archaeological excavation, and relates to an environment perception system, in particular to a binocular vision environment detection system and method based on tactile perception. Background technique [0002] At present, there are a large number of ancient buildings in our country that have not been excavated, especially the ancient buildings in a dangerous environment. For example, due to the long-term closed state of the ancient tomb buildings in the ground, the interior of the ancient buildings has formed a unique ecological environment. Corpses and other substances decayed day by day, and the oxygen in the ancient buildings was exhausted. At the same time, various harmful gases, such as carbon dioxide, hydrogen sulfide, sulfur dioxide, and methane, flooded the ancient buildings. Some ancient buildings such as tombs may also have mercury vapor, and the dark and humid environment underground can also breed mic...

AI Technical Summary

Problems solved by technology

[0017] The present invention provides a binocular visual environment detection system based on tactile perception, aiming at the high noise and low resolution images obtained by existing robots in low-light or no-light environments, which cannot effectively extract image edge feature points, and cannot use The stereo matching method of binocular vision obtains the disparity map, so as to construct the three-dimensional information of the robot's surrounding environment. This system divides the high-noise and low-resolution images into adjacent statistical characteristic loss areas, so that it can The shared sides of the adjacent statistical property loss areas form the shortest curved path for the robot to walk, and then use the tactile sensor to collect the surface information of the object based on this shortest curved path to compensate for the statistical property loss area to generate a stable support point, so that the support point-based The binocular vision stereo matching method is used to obtain the disparity map, and then the three-dimensional information of the robot's surrounding environment can be obtained through the disparity map to construct the corresponding environment model in real time to realize detection.

Images