Image data base and inspection robot-based equipment trouble detection method

Abstract

The invention discloses an image data base and inspection robot-based equipment trouble detection method. According to the image data base and inspection robot-based equipment trouble detection method, an image data base is established, equipment is calibrated manually, a visible light camera and a thermal imaging camera which are installed on an inspection robot are used for shooting, and detection and alarming of equipment with abnormal temperature are achieved only through rectification between a real-time visible light image and the corresponding image data base, so that problems of complex and uncontrollable multi-modal image registration, equipment characteristic extraction, mode recognition and the like are converted to be the problem of single-modal image registration, steps are simplified, and the detection accuracy is improved.

Description

technical field [0001] The invention relates to a device fault detection method based on an image database and an inspection robot. Background technique [0002] At present, my country's equipment monitoring mainly relies on manual observation, such as in substations, power plants and other enterprises. There are many human and uncertain factors in this method. Under the conditions of high pressure, ultra-high pressure, bad weather and harsh environment, not only the personal danger is great, but also certain hidden dangers are brought to the operation of the enterprise. Although some domestic enterprises have carried out monitoring research on inspection robot equipment for the power industry, one of the prominent problems they face is the low utilization rate, mainly "seeing without monitoring", that is, the real-time video stream transmitted by the inspection robot is mainly based on Manual monitoring and information storage are the main ones. Because manual monitoring ...

AI Technical Summary

Problems solved by technology

It is difficult to guarantee practical accuracy in each link

For example, equipment feature extraction, due to the wide variety of equipment, various observation angles and observation times, and various weather conditions, it is difficult to obtain accurate equipment features and accurately identify and locate them; multi-modal image registration itself is a research difficulty, although At present, there are algorithms such as maximum alignment algorithm and mutual information algorithm, but it is very difficult to obtain a very high registration effect in practice; moreover, the accuracy of device pattern recognition and positioning itself based on image processing cannot be guaranteed

Images