Transformer substation flame identification algorithm adopting enhanced RGB component features

Abstract

The invention discloses a transformer substation flame identification algorithm adopting enhanced RGB component features. The algorithm comprises the steps of 1, obtaining transformer substation videodata and performing image preprocessing; 2, processing basic RGB component features; 3, processing enhanced RGB component features; 4, performing dynamic detection based on the enhanced RGB componentfeatures; and 5, identifying a flame phenomenon. According to a calculation method, the flame phenomenon in a transformer substation scene can be enhanced through the enhanced RGB component feature processing; when a small amount of light flames occur, identification also can be accurately performed; in combination with the dynamic detection of the enhanced RGB component features and the detection of the basic RGB component features, the influence of an interference source on the algorithm can be effectively avoided; and by timely discovering accidents, the safety of a transformer substationcan be improved.

Description

technical field [0001] The invention belongs to the technical field of substation intelligent monitoring, in particular to a substation flame identification algorithm using enhanced RGB component features. Background technique [0002] The substation is an important place for the production and operation of the power grid. The safe and stable operation of the substation is of great significance to the reliability of the power grid. The application in production and operation continues to deepen and widen. The video surveillance technology adopted in the traditional way provides remote visual monitoring means for remote monitoring of power grid equipment, and provides remote monitoring technical means for monitoring personnel. Although the equipment in the substation already has very high operational safety and reliability, due to the high load and high temperature operation of the equipment during the summer peak every year, insufficient equipment maintenance, equipment agin...

AI Technical Summary

Problems solved by technology

Although the equipment in the substation already has very high operational safety and reliability, due to the high load and high temperature operation of the equipment during the summer peak every year, insufficient equipment maintenance, equipment aging and other reasons, it also leads to Explosions caused by equipment failures in substations occasionally occur, and the economic losses caused by them are also quite large. Therefore, if the accident phenomenon can be discovered in time when the accident occurs, the economic loss caused by the expansion of the accident will be greatly reduced.

[0003] In the process of substation accidents, there are usually phenomena such as flames and smoke, and most substations are outdoor substations. The commonly used smoke detectors cannot be used in outdoor environments. If the infrared thermal imaging camera is used, although abnormalities can be found However, due to the high results, it cannot be popularized and used. Therefore, by using the monitoring camera installed in the substation and using image intelligent analysis technology to realize the timely discovery of the flame phenomenon in the event of an accident in the substation, it has a high practical value

There are many existing flame recognition techniques, including traditional recognition methods such as the use of flame color feature models, flame smoke features, etc. Each method is usually only used for phenomenon recognition in certain fixed scenes, and other recognition methods include The method of machine learning, which needs to collect a large number of flame phenomenon pictures as machine learning training samples, the integrity of the sample data greatly affects the reliability of recognition

Images