Method and system for automatically identifying monitoring picture of transformer substation

Abstract

The invention provides a transformer substation monitoring picture automatic identification method and system. The method comprises the following steps: reading a substation monitoring picture; extracting a PHOG (Pyramid Histogram of Oriented Grades) feature from the monitoring picture; based on the extracted PHOG features, utilizing a KNN algorithm to judge the category of the monitoring picture; reading the xml file for identifying the electrical primitive position information of the same type of offline monitoring picture, and determining the electrical primitive position and the ID number of the monitoring picture; carrying out binarization processing on the electrical primitives with the determined positions and ID numbers; and when the electrical primitives are in a non-switch type, extracting HOG features from the electrical primitives after binarization processing, and classifying the extracted HOG features by adopting an SVM (Support Vector Machine) to obtain the states of the electrical primitives. According to the invention, automatic identification of the monitoring picture can be realized under the conditions of different display types of monitoring picture operation, electrical primitive color change and the like, the correctness of transformer substation monitoring picture identification is greatly improved, and the efficiency of transformer substation monitoring interface automatic detection is greatly improved at the same time.

Description

technical field [0001] The invention relates to an automatic identification method and system for substation monitoring screens, and belongs to the technical field of operation maintenance and automatic detection of smart grids. Background technique [0002] At present, there are many manufacturers of monitoring system products in substations, with complicated functions and more human-computer interaction processing. According to the current situation, although China Electric Power Research Institute and the testing units of provincial power supply companies have an urgent need for intelligent identification of monitoring screens, there are currently no more comprehensive products developed for automatic identification of substation monitoring screens on the market. According to this situation, this paper adopts methods related to computer vision such as feature extraction and image matching to solve such problems. In recent years, image matching technology has attracted ex...

AI Technical Summary

Problems solved by technology

[0003] At present, there is no relevant research on the direct identification of specific electrical primitives in the monitoring screen. The identification of specific primitives is mainly to determine the category of a single graphic element after preprocessing. For example, by establishing an electrical symbol training set and extracting electrical symbols The HOG feature of the image, calculate the weighted map of the number of gradient direction vectors, use these HOG features and classification information to train the support vector machine, and then use the support vector machine to identify

Images