Research on Automatic Recognition of Multi-class Pavement Disease Based on Convolutional Neural Network
The void ratio and compaction of pavement are the indexes that need to be strictly controlled in the construction of highway,which will greatly affect the performance of highway in the future if it fails to meet the requirements.In order to explore the best method of different pavement void ratio calculation methods,this paper compares three commonly used pavement void ratio calculation methods of ground-penetrating radar,CT scanning and indoor testing.Firstly,the clearance rate is calculated by using the radar data collected from the test section combined with the ALL model,secondly,the pavement core samples are scanned by X-ray CT scanning,and in addition,the clearance rate is calculated by conducting indoor net-basket test on the core samples.By comparing the calculation error values of the three methods,the best method for calculating the pavement void ratio was selected.Through the test,it is concluded that the average relative error between the CT tomography method and the net basket method is 7%,the average relative error between the ground-penetrating radar method and the net basket method is 12%.The ground-penetrating radar method has a slightly larger error than the other two methods,but it is fast,efficient,and economical,and is suitable for large-scale census.
void ratio testing methodsground-penetrating radar(GPR)X-raynon-contactlaboratory testing
万方数据
维普
