Non-destructive Detection System Of Pavement Diseases Based On Andriod Platform

With the continuous increase of highway mileage in China,the occurrence of highway diseases has become more frequent.Nowadays,in order to meet the rigid needs of highway inspection and promote the continuous upgrading of detection methods,the traditional manual detection and semi-automatic detection mode has gradually transformed into an automated and intelligent non-destructive testing mode,among which the detection method based on digital image processing is popular because of its high accuracy and fast speed.However,at this stage,there are still key problems such as excessive reliance on hardware devices,low convenience,and cumbersome data transmission.Therefore,based on image processing technology and computer vision technology,this thesis develops an APP that can realize nondestructive detection of road diseases on mobile terminals,and the main research contents are as follows:(1)The algorithm principle of a variety of image preprocessing methods is introduced,and the analysis is carried out from multiple angles such as realization effect,grayscale histogram,algorithm logic,etc.,and a set of preprocessing algorithms that are most suitable for highway disease images is summarized.At the same time,an optimization algorithm based on the Laplace operator is proposed,which mainly uses the high noise reduction of median filtering on pepper and salt noise,so as to improve the extraction effect of the Laplace operator on the disease profile.The results show that the image processing effect of the proposed algorithm in the highway field is significantly better than that of the traditional Laplace operator and LOG algorithm.(2)A set of detection algorithms for extracting disease feature parameters has been built by combining professional technologies such as pixel capture and global scanning.This set of algorithms fits the functional relationship between the shooting height of the experimental machine and the real area,designs the extraction algorithm of the pavement disease type eigenvalue,proposes the disease type judgment algorithm by combining the rectangle degree,Aspect ratio and projection ratio,and summarizes the analysis algorithm of the disease damage degree according to the industry specifications.By utilizing the Android studio development platform and combining mature technologies such as Opencv library and Gaode Map SDK,the highway disease non-destructive testing APP has been transformed into results,achieving main functions such as disease image acquisition,disease image processing,disease feature extraction,disease spatial positioning,and disease data sharing.And the functional interface has been optimized and designed to improve the user friendliness of this software.(3)The software was functionally tested on two experimental sections,and the results showed that the accuracy of extracting geometric parameters reached over 80%.The recall rate for disease type judgment was 96.38%,the accuracy rate was 95.38%,and the accuracy rate was 97.28%.The F1 score index was 96.14%,which meets the specifications of intelligent testing technology and proves that the APP has certain practicality and credibility.And an error analysis was conducted on the unsatisfactory feature parameters during the detection process,indicating the direction for algorithm improvement.