Aero-engine damage video detection method based on dual space distortion

Abstract

The invention discloses an aero-engine damage video detection method based on dual space distortion. The method comprises the following steps: step 1, determining a key frame in a borescope video; if the current frame is a key frame, executing the step 2, and if the current frame is a non-key frame, executing the step 3; 2, extracting a multi-scale semantic feature map of the current video frame; 3, calculating an optical flow field between the current frame and the previous frame to obtain a distorted semantic feature map; if the current frame is the next frame of the key frame, executing the step 5, otherwise, executing the step 4; step 4, obtaining a distorted semantic feature map of the current frame by calculating a distorted optical flow field of a continuous frame pair between the key frame and the current frame; and 5, performing feature decoding on the multi-scale semantic feature map obtained in the step 2 and the distorted semantic feature map obtained in the step 3 or the step 4 to obtain a semantic segmentation map of the current frame. According to the method, the double-flow field is used for distorting key frame features, and the distortion error of a fast moving object in a borescope video can be solved.

Description

technical field [0001] The invention belongs to the field of computer vision, and in particular relates to a video semantic segmentation method with distorted key frame spatial features, and is a new video semantic segmentation method for quickly detecting damage in an application scene of an aeroengine borehole detection technology. Background technique [0002] Semantic segmentation techniques generate pixel-level lesion-predictive images, in other words, each pixel is classified as lesioned or non-lesioned. The precise lesion location and structure resulting from lesion segmentation can be used both to classify lesion types and to obtain important lesion features. When applied to the field of borehole detection video, due to the complex structure of the image semantic segmentation network, the use of frame-by-frame analysis for semantic segmentation prediction will result in a huge amount of calculation, which cannot meet the real-time performance of detection. [0003] ...

AI Technical Summary

Problems solved by technology

[0004] However, this method is not very suitable for fast-moving objects

This is because there is a certain interval between the key frame and the current frame, and the content of fast-moving objects will change greatly, and the method of single space warping tends to miss the semantic information contained in it, resulting in large distortion errors

[0005] In the application of actual aero-engine borehole detection technology, the internal structure of the engine is complex, and the shooting angle of the borehole detection varies greatly, resulting in short damage occurrence time, fast moving speed and obvious feature changes. Therefore, a single space distortion is not suitable for the actual hole detection technology.

Images