Non-stop inspection method for inspection robot

Abstract

The invention provides a non-stop inspection method for an inspection robot. The non-stop inspection method comprises the following steps: performing target detection on each frame in a video stream in which a target object is shot to obtain a series of qualified frames; calculating the variance of the second derivative of the image for the real-time image of the shot video stream to obtain the fuzzy degree of the image, and adjusting the focusing parameter of the holder; blur = E (delta G-E (delta G)) 2, and taking a Laplacian operator as a convolution kernel to carry out convolution on the image to obtain a second derivative; setting a Blur threshold value, adjusting the focusing parameter from small to large, adjusting the focusing parameter to the Blur threshold value, and outputting the focusing parameter; image depth features are extracted from a series of qualified frames, feature matching is carried out on the qualified frames and the template drawing, the ambiguity Blur is calculated, and when the number of matched feature points is larger than 15, the Score is larger than 4, and the Blur is larger than 500, the image is judged to be the optimal collected image.

Description

Technical field [0001] The present invention relates to a patrol inspection method for inspection robots without stopping. Background [0002] At present, in the business scenario of mobile robot patrol patrol, by pre-planning the inspection path, deploying the relative position of the target, allowing the robot to reach the preset coordinates, stop, perform the inspection task, and then navigate to the next task point after completion, there are the following problems: [0003] (1) It is necessary to set the inspection path and inspection point in advance, in the substation and public security scenarios, the number of inspection points is tens of thousands, and the deployment of inspection points is a very time-consuming and labor-consuming work, which increases the cost of inspection operation and maintenance; [0004] (2) For each inspection task point, the robot needs to stop, shoot the target at a fixed point, and the efficiency and fluency of task execution are very limited...

AI Technical Summary

Problems solved by technology

[0003] (1) It is necessary to pre-set the inspection path and inspection points. In the substation and public security scenarios, there are thousands of inspection points. Deploying inspection points is a very time-consuming and manpower-consuming task, which increases the number of inspections. Operation and maintenance costs;

[0004] (2) For each inspection task point, the robot needs to stop and shoot the target at a fixed point, and the efficiency and smoothness of task execution are very limited;

[0005] (3) In the deployment phase, the preset inspection point is relative to the coordinate position of the inspection robot. Due to the positioning error of the robot and the movement error of the gimbal to adjust the pose, the reliability of shooting the target through the relative position is low, and usually In the event that the target object cannot be photographed or not fully photographed, it will affect the quality of the final inspection;

[0006] (4) Fixed-point shooting is inevitably affected by changes in ambient light, and there is no flexibility to adjust the angle of shooting

Images