Method, system and equipment for acquiring corrosion information of fan blade, and storage medium

Abstract

The invention provides a method, a system and a device for obtaining corrosion information of fan blades, and a storage medium, and the method comprises the steps: S1, obtaining an image set which comprises a plurality of images containing blades; S2, extracting image information of the image according to a predetermined step length and a predetermined direction by using the sliding window until the sliding window traverses the whole image; S3, carrying out classification processing on the extracted image information to obtain image information with corrosion content; and S4, repeating the steps S2 to S3 on the unprocessed image until all image information which meets the preset size requirement and has the corrosion content is obtained. According to the method, system and equipment for obtaining the corrosion information of the fan blade and the storage medium, the corrosion types can be accurately classified, the early-stage manual labeling work is reduced, the algorithm running speed is high, the requirement for the performance of a computer is low, and the calculated amount is far lower than that of a target detection mode in the prior art.

Description

technical field [0001] The present disclosure relates to the technical field of graphic information, and in particular to a method, system, device and storage medium for obtaining erosion information of fan blades. Background technique [0002] At present, the method of target recognition is directly used to identify corrosion, but there are disadvantages: the current target recognition has strict requirements on the image size (the image size needs to be small, 200*300), and the corrosion form on the blade is long strip , the resulting image size is very large, far exceeding the image size requirements for target recognition. Although it is also possible to use the multi-scale division method for target recognition, it is time-consuming, has a huge amount of calculation, requires very high computer performance (processing power of the graphics card), and consumes a lot of computer memory. [0003] Target recognition also requires a lot of time to carry out manual labeling ...

AI Technical Summary

Problems solved by technology

[0004] In addition, the detection ability of target detection for blade corrosion is weak, and it is not enough to distinguish the corrosion types with subdivision differences.

Images