Unlock instant, AI-driven research and patent intelligence for your innovation.

Power defect detection method based on reinforcement learning and Transform

A technology of reinforcement learning and electricity, applied in the field of deep learning, can solve problems such as over-reliance on prior frames, occupation of computing resources, and poor versatility, and achieve the effect of avoiding pooling loss information, reducing the proportion of calculations, and reducing the impact

Pending Publication Date: 2022-07-29
HEFEI UNIV OF TECH +1
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] Aiming at the deficiencies of the existing technology, the present invention aims to solve the existing problems of poor versatility in different environments, excessive reliance on the prior frame, irrelevant backgrounds occupying computing resources and affecting the accuracy of the existing power defect detection methods, and proposes a depth-based Power defect detection method, in order to realize intelligent power defect inspection, reduce labor costs, and improve detection efficiency and detection accuracy

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Power defect detection method based on reinforcement learning and Transform
  • Power defect detection method based on reinforcement learning and Transform
  • Power defect detection method based on reinforcement learning and Transform

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0053] In this embodiment, refer to figure 1 , a power defect identification method based on reinforcement learning and Transformer is carried out as follows:

[0054] Step 1. Collect the power inspection aerial image set and use the deep convolutional generative adversarial network for data enhancement to obtain the expanded image data set. The image is marked with a target detection frame, thereby obtaining a training data set; in this embodiment, there are a total of 8192 pictures in the training data set;

[0055] Step 2. Build a reinforcement learning module for screening the foreground region and background region, and input the training data set into the reinforcement learning module for training, and obtain the foreground region feature vector set F of the ith image in the training data set i,f and background area vector set F i,b ;

[0056] Step 2.1, build a reinforcement learning module, including: the backbone network and DQN network for DetNet feature extraction...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses an electric power defect identification method based on reinforcement learning and Transform, and the method comprises the steps: 1, carrying out the aerial photographing of an unmanned plane, collecting an original data set, and generating an adversarial network augmentation data set through deep convolution; 2, extracting image features by using a reinforcement learning module to search a foreground region; and 3, compressing the feature vector of the background region through a Transform module, further performing feature extraction, and finally obtaining a final prediction result through a full connection layer. According to the method, deep learning is utilized to realize detection of the electric power defect area, so that the labor cost is reduced, the method is not influenced by external factors such as weather and background, and the detection efficiency and the detection precision are improved.

Description

technical field [0001] The invention relates to the field of deep learning, in particular to a power defect detection method based on reinforcement learning and Transformer. Background technique [0002] With the continuous growth of power demand in my country and the rapid expansion of the scale of transmission lines, transmission lines have been exposed to complex natural environments for a long time, and are easily disturbed by natural conditions such as rainwater corrosion and lightning strikes, which are prone to various faults and affect the safe and stable operation of the power grid. Therefore, high-efficiency inspection of power transmission lines is essential to ensure the normal operation of production and life. [0003] Many transmission lines are erected in inaccessible areas with complex terrain, and traditional manual inspections are costly and difficult. Therefore, in many areas, drones are equipped with image acquisition devices to quickly collect images an...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06V20/17G06N3/04G06N3/08G06V10/764G06V10/774G06V20/40
CPCG06V20/17G06V20/41G06V10/765G06V10/774G06V20/46G06N3/08G06N3/045Y04S10/50
Inventor 李帷韬侯建平胡平路管树志杨盛世张雪松李奇越孙伟刘鑫常文婧李卫国王刘芳董翔宇黄杰
Owner HEFEI UNIV OF TECH