Method for quickly detecting power transmission line channel hidden dangers

Abstract

The invention discloses a method for quickly detecting hidden dangers of a power transmission line channel based on space segmentation and multi-feature matching. The method comprises the following steps: firstly, regularly acquiring a field image, dividing the acquired image into a sky area and a ground area, then judging hidden dangers of the large construction machinery in the sky area image byadopting a feature analysis method, and if the hidden dangers exist, transmitting the current image to a monitoring center for accurate analysis; selecting a reference image for the ground area afterthe sky area is removed by utilizing weather and moment information; and finally, carrying out multi-feature extraction and decision-level fusion to judge whether the ground area has hidden dangers or not. According to the invention, the characteristics of power transmission line channel images and hidden dangers are deeply excavated; a space segmentation and multi-feature matching method is adopted to quickly detect hidden dangers of a power transmission line channel, the defects that a traditional universal algorithm is low in calculation speed, low in accuracy and the like are overcome, meanwhile, all monitoring images are prevented from being transmitted to a monitoring center, the power consumption and the flow consumption of the device are reduced, and the availability of the deviceis increased.

Description

technical field [0001] The invention relates to a method for quickly detecting hidden dangers in transmission line passages, and belongs to the technical field of transmission lines. Background technique [0002] In recent years, the scale of the power grid has expanded rapidly, and the pressure and challenges faced by the operation and maintenance of transmission lines have also increased. Large-scale construction machinery such as excavators, cranes, and tower cranes have become one of the main reasons for line trips. At present, the existing transmission line protection schemes mainly rely on the transmission line channel visual monitoring device to regularly acquire line channel images, and then transmit them to the background server through the 4G network. The background server uses artificial intelligence image recognition methods to identify hidden dangers. However, the monitoring device deployed on the transmission tower can only be powered independently by solar ene...

AI Technical Summary

Problems solved by technology

However, due to the influence of rainy weather, the texture features at the junction of the sky and the ground are easily covered up

Document 2 (Yu Peixiang, Wu Zhong, etc., "An Efficient Online Monitoring Algorithm for Electric Power Image", East China Electric Power, 2013). Using the image in the stable period as the background, calculate the difference between the real-time image and the current image, and use the maximum probability criterion to determine the abnormal situation .But this method is only applicable to substations with a single scene

Document 3 (Li Chengqi, Ren Zhigang, etc., "Binary image filtering for object detection based on Haar feature density map", Proc. SPIE 10613 1061303) proposed a binary image filtering method based on Haar feature density map, which can eliminate the noise caused by Small disturbances, but it is not clear whether the large connectivity differences are caused by human factors or natural factors

[0004] The above method has problems such as single recognition target and single recognition scene. It does not combine the characteristics of the transmission line channel, and has done too much analysis on the nature of the transmission line channel scene itself, so the adaptation is poor and the accuracy rate is low.

Images