Video-based intelligent building inspection method and system

Abstract

The invention provides a video-based intelligent building inspection method and system. The method comprises the steps of obtaining a historical building video and a real-time building video, obtaining training data from the historical building video, carrying out the processing and marking of the training data, and obtaining a marking data set, constructing a training neural network, and processing the real-time building video through the training model to obtain a to-be-tested device category, a historical inspection record of the to-be-tested device and a relative coordinate of the to-be-tested device in the image, obtaining an inspection layout scheme according to the device category, the historical inspection record and the relative coordinates, wherein the inspection layout scheme specifically comprises planning an inspection route and inspection time and formulating an inspection task table, and effectively judging the relation between the target entity area of the image and the adjacent device, line and environment scenes, effectively reducing the inspection error, and meeting the requirement for achieving building inspection accuracy through videos.

Description

technical field [0001] The present disclosure relates to the field of building inspection and maintenance, and relates to a video-based intelligent building inspection method and system. Background technique [0002] The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art. [0003] Electrical maintenance and inspection in buildings, especially large buildings, is an important task. In order to prevent the safety of electrical facilities, electrical facilities are set in special control cabinets, but for the maintenance of control cabinet doors and electrical facilities, through Without intelligent management, the cabinet door is simply locked, and repair reports and repairs are also transmitted through manual oral or paper information, which is inefficient and cannot store and record information in a timely manner. , especially for the switch control of the cabinet door during the ma...

AI Technical Summary

Problems solved by technology

[0006] In the traditional inspection management work, the methods of listing, dialing needles, and signing in are generally used. The disadvantages of these methods are that early inspections, late inspections, missed inspections, or even no inspections often occur without inspections in accordance with inspection regulations. If the manager does not visit the site in person, it is impossible to grasp the inspection situation of the inspectors, and it is impossible to know whether the inspectors are conducting inspections on time and according to the specified inspection plan.

In addition, the workload and efficiency of personnel patrol inspection are large. Due to the large variety of manual records and heavy workload, the work efficiency of patrol inspection and reporting is low, and missing or wrong items occur from time to time. Because the description of the same defect is not uniform , there are often problems such as duplication of records, difficulty in sorting, inconvenient file preservation and future search, and cannot provide a scientific and technical basis for defect elimination

[0007] Through the automatic inspection of the video surveillance network, the cameras in the network can be inspected in a timely and rapid manner. In the prior art, the video inspection server generally performs inspections in the alarm center, which is inconvenient to deploy at the network point and the speed is slow. And easily affected by network transmission speed;

[0008] Moreover, the problem of building inspection is often related to the target entity in the video. However, the information extracted from the current video cannot realize the extraction of shape information, and cannot effectively judge the target entity area of ​​the image and the adjacent equipment, lines and environmental scenes. The relationship between them makes the inspection error relatively large. At present, it is difficult to meet the needs of realizing the accuracy of building inspection through video.

Images