Safe anti-violation video automatic sorting system model based on computer vision

Abstract

The invention discloses a safe anti-violation video automatic sorting system model based on computer vision, and belongs to the technical field of computer sorting. Safety anti-violation video automatic sorting is carried out through a basic support layer, a technology layer, an application layer and a service layer. The production efficiency and the automation degree of production can be greatly improved by utilizing a machine vision detection method. The application of a machine vision product has an obvious cost advantage on the substitutability of manual work, and has a higher consistency requirement. The saliency map generated by the model provided by the invention not only can effectively cover the boundary of the power equipment, but also has high contrast ratio between the equipment area and the background area, so that the saliency of the background area is well inhibited. According to the method, deep learning is closely related to a traditional machine learning method, a deep learning module is added into a traditional machine learning framework, and higher equipment sorting recognition accuracy can be obtained.

Description

Technical field [0001] The present invention belongs to the field of computer sorting technology, specifically relates to a secure anti-violation video automatic sorting system model based on computer vision. Background [0002] In order to improve the level of intelligence of the power system, many substations have not only the four remote functions of "telemetry", "remote communication", "remote control" and "remote adjustment", but also added the "remote vision" function, using remotely transmitted images and videos to timely evaluate the operation of related equipment. The introduction of image and video information into the power system is a huge technological progress, however, the existing image and video surveillance system only transmits a large amount of media data to the dispatching end, can not intelligently analyze these data, and requires staff to manually analyze and identify the received data. The manual processing method is inefficient, the accuracy rate is not h...

AI Technical Summary

Problems solved by technology

From the perspective of actual application status, the technical level of power video surveillance systems used in different business scenarios is uneven. Most substations are equipped with video surveillance systems, but the systems are independent of each other and cannot achieve information sharing. From the perspective of video surveillance technology, each There are also big differences in the technology used by manufacturers, which cannot meet the actual needs of the management of the power industry, especially the large amount of video recording. During the inspection, the management personnel cannot confirm the key points of the key links. They can only view and compare all the recorded videos one by one. Sorting, the workload is huge, the work content is repetitive, and it takes a lot of time to complete the videos and information of key links and key items

Images