PCA-based end-to-end pavement crack detection-recognition method

Abstract

The invention discloses a PCA-based end-to-end pavement crack detection-recognition method and relates to the field of computer vision. The method comprises the steps that S1, a dataset about a pavement image I<x> is acquired, and the dataset is preprocessed; S2, the preprocessed dataset is subjected to category marking, a background image I not containing a crack is calculated, a difference image I between the pavement image I<x> and the background image I is acquired, and a training set and a test set are constructed according to proportion; S3, training set images I are utilized to train a deep neural network model; S4, test set images I are input into the trained deep neural network model to acquire image category probabilities; and S5, a category tag corresponding to the maximumprobability is selected to serve as an image detection-recognition result. Through the scheme, traditional detection and recognition processes are fused, and therefore the efficiency and robustness of pavement crack detection are improved.

Description

technical field [0001] The invention relates to the field of computer vision, in particular to a PCA-based end-to-end pavement crack detection and recognition method. Background technique [0002] Whether it is asphalt or cement pavement, after a period of time of traffic opening, cracks appear on the pavement due to the external environment, which brings great hidden dangers to the normal use of the pavement. Therefore, an effective detection and evaluation method is needed to detect and identify possible hidden dangers, so as to avoid potential harm. Pavement crack detection refers to the process of first detecting whether there are cracks or defective parts in the road surface in the image, and then identifying the crack type. It involves two processes of detection and recognition. This is a key issue in computer vision. It is based on image The basic technology of content recognition. Pavement crack detection can be used in road restoration, road condition monitoring a...

AI Technical Summary

Problems solved by technology

However, due to the differences in road conditions and inconsistencies in various scenarios, this detection method is inefficient and has large errors or even fails, and it is impossible to determine the type of cracks.

Images