Substation inspection robot image semantic analysis system and method

Abstract

The invention discloses a substation inspection robot image semantic analysis system and method. The system comprises an embedded AI analysis module, a binocular vision camera, an inspection camera, and a three-dimensional laser radar multi-degree-of-freedom mechanical arm, wherein the binocular vision camera, the inspection camera and the three-dimensional laser radar multi-degree-of-freedom mechanical arm are connected with the embedded AI analysis module, and the embedded AI analysis module comprises a roaming semantic map construction module; and the embedded AI analysis module is configured to be acquiring binocular image data, inspection image data and three-dimensional point cloud data of a current environment in real time; acquiring spatial distribution of objects in the current environment based on the binocular image data and the three-dimensional point cloud data; and achieving automatic recognition of the passable unknown region around the robot, achieving the motion planning of the robot in the unknown region through a local path planning technology, and executing the map construction of the unknown environment till the construction of the whole in-station environmentsemantic map is completed. According to the system, real-time and efficient identification of the inspection video is realized, and the accuracy of substation equipment identification is improved.

Description

technical field [0001] The invention belongs to the technical field of substation inspection robots, and in particular relates to an image semantic analysis system and method for substation inspection robots. Background technique [0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute prior art. [0003] At present, the inspection points of substation inspection robots are usually manually set by on-site personnel according to the inspection tasks. When setting, the on-site personnel first remotely control the robot to run along the inspection route, and stop when it runs to the periphery of the power equipment to be inspected; Adjust the posture of the robot pan / tilt, so that the pan / tilt drives the visible light camera, infrared thermal imager and other non-contact detection sensors to align with each device to be inspected around the robot in turn and record the corresponding pan / til...

AI Technical Summary

Problems solved by technology

[0004] The above inspection method has at least the following problems: (1) The existing substation inspection robot has weak front-end video and image processing capabilities, and most of the image and video data is currently analyzed and processed by the back-end server through the network, which is affected by the bandwidth of the transmission network , there is a certain delay in data analysis, which cannot meet the needs of application scenarios with high real-time requirements such as robot navigation, visual servoing, and timely defect detection; When docking at a fixed detection point and then collecting inspection data from surrounding equipment in different directions, it is difficult to obtain high-quality inspection data for each device due to the impact of sensor observation angle and light, which is not conducive to subsequent equipment status identification and faults Diagnosis; (3) The setting of inspection points is greatly affected by the subjective judgment of on-site configuration personnel, and the setting standards are inconsistent, which leads to the inability to guarantee the quality of the setting of inspection points; moreover, the manual setting process of inspection points is cumbersome and heavy workload Big

Images