Deep learning-based power transmission and transformation project quality common disease prevention and detection method

A technology of engineering quality and deep learning, applied in the field of image recognition and computer vision, can solve problems such as low efficiency, waste of human resources, and reduce efficiency, and achieve the effect of good generalization ability, robustness, and good detection performance.

CN112598054APending Publication Date: 2021-04-02福建京力信息科技有限公司
0 Cites 1 Cited by

Patent Information

Authority / Receiving Office
CN · China
Current Assignee / Owner
Publication Date
2021-04-02

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention relates to a deep learning-based power transmission and transformation project quality common problem prevention and detection method, which comprises the following steps of S1, obtaining power transmission and transformation project power box grounding connection detection data, and performing preprocessing; S2, constructing a power transmission and transformation project power boxgrounding connection detection data set according to training algorithm requirements; S3, adjusting and optimizing training hyper-parameters of a deep learning algorithm yolov4tiny, optimizing the training model by adopting an optimization algorithm, and performing training according to the data set to obtain a yolov4tiny detection model; S4, performing target detection on the input picture according to the yoov4tiny detection model obtained after training, and obtaining a preliminary detection result; S5, decoding the preliminary detection result, screening out a final detection result by adopting an improved non-maximum suppression algorithm, and finally drawing a detection box in the input picture. The method can effectively identify and judge whether the power box is in grounding connection or not, has good generalization ability and robustness, and can have good detection performance in a complex environment.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention relates to the fields of image recognition and computer vision, in particular to a method for preventing and detecting common quality problems of power transmission and transformation projects based on deep learning. Background technique

[0002] With the increasing importance of the normal operation of the power system in national production and life, the common quality problems in the construction of power transmission and transformation projects have also received high attention. As early as 2010, the State Grid Corporation compiled the "Requirements and Technical Measures for the Prevention and Control of Common Quality Problems in Power Transmission and Transformation Projects of State Grid Corporation" based on the national and industry-related engineering construction quality standards and specifications From the technical point of view, specific prevention and control measures are put forward, and the prevention and control work r...

Examples

Embodiment Construction

[0052] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0053] Please refer to figure 1 , the present invention provides a kind of power transmission and transformation engineering quality common fault prevention and detection method based on deep learning, comprising the following steps:

[0054] Step S1: Obtain the ground connection detection data of the power box of the power transmission and transformation project, and preprocess it;

[0055] Step S2: According to the requirements of the training algorithm, construct the ground connection detection data set of the power box of the power transmission and transformation project;

[0056] Step S3: tune the training hyperparameters of the deep learning algorithm yolov4-tiny, adopt an optimization algorithm to optimize the training model, and train according to the data set and obtain the yolov4-tiny detection model;

[0057]Step S4: according to the yolov4...