Pavement Cracks Automated Processing Based On Image Detection with Neutral Network









Abstract

Pavement crack detection and the width calculation of cracks are important in roads management and maintenance. This paper uses an image processing method after image classification to calculate the crack width. In the image classification procedure, the different color model images that in R, G, B three channels as the inputs in the neutral network and the results of the softmax classifiers can find the input image whether belong to the crack image or not. By comparing and calculating, the accuracy and precision of the image classification can be provided. After that, the outputs of the convolution neutral network will run through the two median filters and a Gaussian filter that are used for moving noise and smoothing images to achieve the clear crack images. The results are provided in the calculation of the pavement crack images.


Modules


Algorithms


Software And Hardware

• Hardware: Processor: i3 ,i5 RAM: 4GB Hard disk: 16 GB • Software: operating System : Windws2000/XP/7/8/10 Anaconda,jupyter,spyder,flask Frontend :-python Backend:- MYSQL