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

Electric power scene video detection method based on environment self-adaption and small sample learning

A video detection and self-adaptive technology, applied in character and pattern recognition, image analysis, image enhancement, etc., can solve the problems of less research, unclear proportion of animals in the monitoring screen, difficult to be recognized, etc., and achieve the effect of overcoming mutation

Pending Publication Date: 2020-06-26
SHANDONG UNIV +3
View PDF4 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the identification and monitoring method in this literature is difficult to apply to the field of substations with changing environmental conditions to detect the intrusion behavior of live animals
[0007] In terms of substation target detection, due to the wide variety of invading animals, and the animals are not clear or occupy too small a proportion in the monitoring screen, it is difficult to be identified, which seriously affects the detection efficiency
In addition, the existing picture and video data of animal invasion are quite scarce, which makes substation animal target detection face more challenges than general target detection tasks.
[0008] Due to the particularity and complexity of animal invasion scenarios in substations, and there are few related studies, there are many difficulties in substation motion detection and target detection.

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
  • Electric power scene video detection method based on environment self-adaption and small sample learning
  • Electric power scene video detection method based on environment self-adaption and small sample learning
  • Electric power scene video detection method based on environment self-adaption and small sample learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0091] Apply the power scene video detection method based on environment adaptation and small sample learning of the present invention to the detection of small animal intrusion in the power scene, specifically including:

[0092] 1) Motion detection:

[0093] The improved generalized Gaussian mixture model (GGMM) is used to perform environment-adaptive motion detection on the surveillance video, and the detected moving objects are marked with candidate frames;

[0094] At the same time, the corresponding image content is cropped according to the candidate frame;

[0095] Then, the cropped "close-up" image is used as the input of the target detection in the next link;

[0096] Finally, receive the result of target detection for visual marking or alerting.

[0097] 2) Target detection:

[0098] First, make targeted adjustments to the YOLOv3 model;

[0099] Then, input the "close-up" image into the improved YOLOv3 model for target detection, and determine whether animal inva...

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 scene video detection method based on environment self-adaption and small sample learning. The method comprises the following steps: motion detection: carrying out environment self-adaption motion detection on a monitoring video by adopting an improved generalized Gaussian mixture model, and marking a detected moving target by using a candidate box; performing corresponding image content cutting according to the candidate box; taking the clipped close-up image as the input of the target detection of the next link; receiving a target detection result, and carrying out visual marking or alarming; target detection: performing targeted adjustment on the YOLOv3 model; and inputting the close-up image into an improved YOLOv3 model for target detection, and judging whether animal invasion occurs or not by judging whether the object is an animal or not. The method has good real-time performance, accuracy and robustness, not only can meet the actual requirements of animal invasion detection of the transformer substation, but also further enriches the research in related fields.

Description

technical field [0001] The invention discloses a power scene video detection method based on environment self-adaptation and small sample learning, and belongs to the technical field of power environment intelligent recognition. Background technique [0002] With the development of society, my country's demand for electric energy is increasing day by day, which promotes the continuous growth of the number of electric equipment. Therefore, strengthening the safety monitoring of electric power scenes is the main preventive measure to ensure the safe and stable operation of electric power. Video surveillance in the early electric power industry mainly played the role of recording, and it was usually dealt with after the accident occurred. In recent years, the reform of the power industry has been advancing, and the detection technology of power scene monitoring video has also achieved rapid development. The video detection method based on computer vision and deep learning can ...

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): G06T7/246G06T5/30G06K9/00G06K9/62
CPCG06T7/246G06T5/30G06T2207/10016G06T2207/20081G06V20/40G06V20/46G06V20/52G06V2201/07G06F18/23
Inventor 聂礼强郑晓云战新刚姚一杨刘晓川刘萌
Owner SHANDONG 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