Image classification method and system for intelligent pump cavity endoscope fault diagnosis

Abstract

According to the image classification method and system for intelligent pump cavity endoscope fault diagnosis, a residual multi-scale feature extraction module is combined structurally, a jump connection mode is adopted, feature integrity is guaranteed, meanwhile, the calculation amount is reduced, and the expression ability of a network model for input image feature information is enhanced; for the problem of low quality of a current scene-oriented endoscope image, an inter-cut super-resolution model is added to improve the quality of the image so as to improve a fault classification result of a low-quality blurred picture; for an input feature matrix, on one hand, redundant feature information is filtered by using a channel attention mechanism, and on the other hand, redundant background feature information is removed by using a space attention mechanism, so that accurate classification is performed.

Description

technical field [0001] The invention belongs to the field of intelligent pump cavity endoscopes, in particular to an image classification method and system for fault diagnosis of intelligent pump cavity endoscopes. Background technique [0002] In the era of rapid development of artificial intelligence, intelligent pumps have been widely used in important fields such as industry, agricultural production, energy, petrochemicals, aviation, steel, and military industry, and play a very important role in the development of the national economy. As a big manufacturing country, my country's smart pump manufacturing technology still has many problems, such as the failure of smart pumps caused by insufficient investment in research and development, weak independent innovation capabilities, and weak basic supporting components. Corrosion, rust, and other internal faults in the pump chamber are often difficult to diagnose by human eyes. As a result, the service life of the smart pump i...

AI Technical Summary

Problems solved by technology

[0005] In the current industrial smart pumps, it is difficult for the human eye to directly observe the internal defects of the pump cavity. After using the endoscope to obtain the detection image, it is difficult to accurately extract the discriminative regional features of the image by image classification. The existing residual network algorithm There are still deficiencies in feature extraction, and the classification accuracy still needs to be improved; in the process of endoscopic detection image classification, the network's ability to express image features is insufficient

Images