A power grid equipment image weak supervision instance segmentation method based on RGB-T fusion

Abstract

The invention discloses a power grid equipment image weak supervision instance segmentation method based on RGB-T fusion, which comprises the following steps: 1, acquiring RGB images and infrared thermal images of pairs of a plurality of power grid equipment, and processing the acquired RGB images to obtain simple RGB images; 2, automatically marking a mask on the simple RGB image; 3, establishingan instance segmentation model; Constructing a simple training set to train the instance segmentation model; 4, performing segmentation prediction on the original RGB image by using the trained instance segmentation model to obtain a segmentation mask mark of the original RGB image; Constructing a complex training set to train the instance segmentation model to obtain a final power grid equipmentinstance segmentation model; And 5, acquiring an RGB image of the power grid equipment, and inputting the RGB image into the final power grid equipment instance segmentation model to obtain a power grid equipment segmentation mask. According to the method, automatic marking of samples can be realized, so that quick and accurate power grid equipment instance segmentation is realized.

Description

technical field [0001] The invention relates to the field of image instance segmentation, in particular to a method for image instance segmentation of grid equipment with weak supervision based on RGB-T fusion. Background technique [0002] Intelligent image recognition and diagnosis technology is an online monitoring technology developed with the development of artificial intelligence research and the leap of photography and imaging technology. Real-time detection and identification of power grid equipment by using inspection robots and portable smart devices can reduce on-site operation and maintenance. The workload of personnel is effectively guaranteed to ensure the safe and reliable operation of equipment. Deep learning technology automatically learns more effective image feature expression from massive data, and is successfully used for image detection and segmentation of general objects. However, there are still huge difficulties and challenges in applying deep learni...

AI Technical Summary

Problems solved by technology

There is a large difference in resolution between infrared thermal images and RGB images, which easily leads to loss of RGB image information, as well as data loss of certain scales and scenes, which leads to insufficient model training

Images