Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Method for identifying insulator defects by pruning YOLOv3 with small on-site sample size

A technology of insulators and sample size, applied in neural learning methods, character and pattern recognition, optical test flaws/defects, etc., can solve the problems of missed judgment and misjudgment of manual inspection pictures, inability to always use learning results, and cumbersome screening work. , to achieve the effect of optimizing lightweight performance, avoiding precision degradation, and reducing hardware requirements

Pending Publication Date: 2020-01-17
SOUTHEAST UNIV
View PDF1 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although this method solves the danger of manual inspection, the screening work in the later stage is still cumbersome and complicated, and manual inspection of pictures due to problems such as picture quality and clarity will cause many missed and misjudged results
Later, the artificial intelligence algorithm based on the deep learning model was gradually used for image recognition and screening. However, different inspection sites have different characteristics, and the same set of learning results cannot always be used. The small number of on-site samples and fewer effective samples have become unavoidable. question

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
  • Method for identifying insulator defects by pruning YOLOv3 with small on-site sample size
  • Method for identifying insulator defects by pruning YOLOv3 with small on-site sample size
  • Method for identifying insulator defects by pruning YOLOv3 with small on-site sample size

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0032] The present invention will be further explained below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention.

[0033] As shown in the figure, a YOLOv3 pruning method suitable for identifying insulator defects with a small sample size on site, the specific steps are:

[0034] Step 1: According to the real-time transmission of image data according to the drone inspection, obtain the insulator image and the corresponding XML label data, and change it to TXT format according to the file characteristics of Darknet-53. The labeling tool LabelImg used is a visual image calibration Tool, the data set required by the YOLOv3 algorithm needs to use this tool to calibrate the target in the image and generate an XML file that follows the format of PASCAL VOC;

[0035] Step 2: Carry out pre...

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 provides a method for identifying insulator defects by pruning YOLOv3 with a small on-site sample size. The method comprises the steps: 1, transmitting picture data in real time according to the inspection of an unmanned plane, and obtaining an insulator image and corresponding XML label data; 2, processing the insulator image; 3, performing pruning modification operation by taking YOLOv3 as a basic framework, and then establishing a deep learning model; 4, introducing an SPPnet module composed of four parallel maxpool layers into the pruned YOLOv3, and enabling the SPPnet moduleand the pruned YOLOv3 to be subjected to pruning; 5, converting the processed insulator image and the modified TXT label data into training data; 6, obtaining an insulator recognition model; 7, obtaining insulator picture frame information detected in each picture; and 8, judging the fault of the insulator. The method has the advantages that under the condition that the average precision is not reduced, the training time is shortened, and the requirement for the picture quality is reduced, meanwhile, the requirement for the computing power of a vehicle-mounted server is reduced, and the intelligent level of insulator inspection is improved.

Description

technical field [0001] The invention relates to power grid inspection and maintenance technology, in particular to a YOLOv3 pruning method for identifying insulator defects with a small number of on-site samples. Background technique [0002] With the rapid development of the economy and the increasing progress of smart grid construction, higher requirements are put forward for the intelligent and rapid inspection of transmission lines. As we all know, an insulator is a special insulation control that can play an important role in overhead transmission lines. Its main function is to achieve electrical insulation and mechanical fixation, for which various electrical and mechanical performance requirements are stipulated. However, the process of manual maintenance and overhaul is also extremely cumbersome and has a high risk factor. According to survey statistics, there are many safety accidents caused by insulator defects. Therefore, regular inspection of tower insulators is...

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
IPC IPC(8): G06K9/00G06K9/62G06N3/08G06Q50/06G07C1/20G01N21/95
CPCG06Q50/06G07C1/20G06N3/082G01N21/95G06V20/10G06F18/214
Inventor 陈嘉琛俞曜辰陈雨辰顾雅茹陈中
Owner SOUTHEAST UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products