Pipeline defect identification method

Abstract

The invention provides a pipeline defect identification method. A pipeline robot transmits a video in a pipeline in real time, samples the video into a key frame image, effectively identifies and captures pipeline defects in the image, and stores an extracted defect image positioning mark. According to the method, the basic idea of DefectNet is used for solving the target defect detection as a regression problem. Firstly, feature extraction is performed through a convolutional neural network to obtain a feature map of a certain size; the then, feature maps of different scales are divided intoa plurality of grids by multi-scale prediction, and the grid where the center of a defect target is located is responsible for predicting the defect target; and finally, target classification and frame regression are carried out, and each grid judges the category to which the defect target belongs and adjusts the frame position. Compared with the prior art, the method has the advantages that the detection precision and the detection efficiency are improved, and the method has high innovativeness and practical value and is suitable for popularization.

Description

technical field [0001] The invention belongs to the field of computer vision and machine learning, and in particular relates to a pipeline defect identification method. Background technique [0002] Traditional pipeline defect identification methods use manual real-time detection, or use traditional typical defect detection algorithms Haar feature + Adaboost algorithm, Hog feature + Svm algorithm, DPM algorithm, etc. for defect detection. The workload of manual detection is heavy, and there are still some problems in the traditional algorithm: the traditional defect detection algorithm uses sliding windows to select candidate areas without pertinence, and at the same time, when using sliding windows to traverse a picture, all windows need to be calculated separately , with high time complexity and redundant windows. This seriously affects the efficiency of subsequent feature extraction and classification. The size of the window needs to be set manually, and different scale...

AI Technical Summary

Problems solved by technology

[0003] In order to solve the defects existing in the above-mentioned background technology, a pipeline defect recognition method provided by the present invention uses the Defect_Net intelligent detection model to perform target detection, which solves the problem of low classification efficiency and low accuracy in the traditional defect recognition algorithm in the current technology , is applicable to the detection of defects on the inner wall of the pipeline. In order to achieve the above purpose, the technical solution of the present invention is as follows:

Images