Research On 3D Detection Method Of Pavement Texture Depth Based On Laser Speckle

With the rapid development of China’s highway transportation infrastructure,China’s highway industry has shifted from infrastructure to construction and maintenance,and accurate and effective detection of pavement is the guarantee for the realization of traffic informatization.The texture depth is important in pavement detection Indicators are important parameters of pavement anti-skid and drainage,therefore,it is necessary to scientifically and accurately detect the texture depth to evaluate its road construction and operation,and guide its maintenance and maintenance work.Due to the defects in the measurement,the existing measurement methods can no longer meet the needs of automated and high-precision detection.Therefore,research and development of high-precision and high-efficiency automatic real-time detection methods and devices for the texture depth structures are imminent,which have an important role in China’s road maintenance detection.Based on the analysis of traditional texture depth detection methods and equipment performance characteristics,combined with the principle of laser speckle imaging,speckle characteristics and influencing factors,a three-dimensional texture depth detection method and device based on laser speckle is designed.To improve accuracy and real-time as the detection target,focusing on the key factors affecting the quality of asphalt speckle image research,principle prototype design,digital speckle correlation algorithm(DSCM)optimization,speckle-related 3D reconstruction model,3D texture depth calculation method and evaluation indicators,etc.The technology was researched,and the principle prototype of the texture depth three-dimensional detection system suitable for outdoor conditions was finally designed,and the rationality and effectiveness of the system were verified through experiments.This paper first theoretically analyzes the first-order and second-order characteristics of laser speckle and its mathematical model,conducts a simulation experiment of simulated speckle model based on its characteristics,According to the global evaluation parameter average gray gradient and the local evaluation parameter sub-region entropy,the quality influence factors of the asphalt material speckle image were tested,and the principle prototype of pavement speckle detection was designed.Then,the digital speckle correlation algorithm was improved and optimized,the cross search method based on the relevant area and the sub-pixel method based on six-point binary quadratic polynomial surface fitting are proposed.The search window size in the algorithm is compared and the optimal window size is 50?50.And according to the relationship between the offset obtained by the digital speckle correlation method and the two-dimensional coordinate,a three-dimensional model of correlation matching is established.For the edge burr problem in the three-dimensional reconstruction,the corrosion expansion algorithm and the mean optimization algorithm are proposed to eliminates burrs on the edges and improves measurement accuracy.Secondly,the calculation model and method of the three-dimensional texture depth are given.The three-dimensional data is correlated with data of manual sanding method to obtain the relationship between the two and the difference between the SMTD parameters is given.The range is used as the depth evaluation index of the three-dimensional detection structure.Finally,by detecting the pavement in different conditions and comparing it with the sanding data results,it shows that the three-dimensional data measurement accuracy is above 89.5%.The experimental results show that the method can accurately the detection of the texture depth of asphalt pavement has great practical value.