Cruising robot pan-tilt adjustment method based on visual feedback

Abstract

The invention discloses a cruising robot pan-tilt adjustment method based on visual feedback. The method comprises the steps that a cruising robot conducts cruising and transmits a real-time shot and collected equipment image to a server in the cruising process; the server matches the real-time collected equipment image with matching templates stored in a database; the position offset relation between the real-time collected equipment image and a template image and the matching value of the two images are computed; the completeness of the collected equipment image is judged according to the computed matching value; when it is judged that the collected equipment image is incomplete, pan-tilt posture data and camera working parameters of the cruising robot are adjusted, and the cruising robot is controlled to conduct cruising on a next preset point when it is judged that the collected equipment image is complete or the number of adjustment times reaches a preset time number threshold value. By means of the method, automatic pan-tilt adjustment of the cruising robot can be realized, the control precision and intelligent degree are high, operation is convenient and fast, and the method can be widely applied to the industry of cruising robots.

Description

technical field [0001] The invention relates to the field of automatic control, in particular to a method for adjusting the pan-tilt of a patrol robot based on visual feedback. Background technique [0002] Glossary: [0003] ROI: region of interest, region of interest. [0004] With the pace of science and technology, the construction of the power grid is developing from the traditional manned manual inspection to the unattended intelligent substation. Therefore, the inspection robot was born. The inspection robot system is a complex system integrating information integration technology, multi-sensor fusion technology, autonomous navigation technology, security technology, advanced image processing technology, pattern recognition technology, and efficient wireless transmission technology. Based on laser autonomous navigation, the patrol robot realizes the optimal planning of the inspection route, automatically avoids obstacles and detours, collects images, audio, and infr...

AI Technical Summary

Problems solved by technology

This kind of terrain environment will inevitably have a certain impact on the driving wheels of the robot, which leads to a certain error in the repeatability of the parking position at the preset position during the inspection process. At this time, the heading angle of the robot is relatively There will be differences when the template image is collected. Such a difference will lead to deviation errors in the equipment area in the collected image even if the robot accurately controls the attitude of the gimbal.

[0008] (3) In an ordinary 220kV substation, the robot needs to observe more than 3,000 pieces of equipment when performing inspection tasks, and the pan-tilt of the inspection robot has a load of about 5kg. After a certain period of use, the pan-tilt rotates tens of thousands of times The mechanical equipment of the gimbal will inevitably wear out. If the same angle rotation control parameters are used all the time, the actual attitude of the gimbal may change, that is, the device image may shift out of the field of view of the image.

[0009] (4) When the gimbal rotation control is no lon

Images