Study On Discrimination Method Of Asphalt Pavement Based On Digital Image

As China enters a new stage of rapid development,policies such as a strong transportation country,infrastructure is continuously improved,road traffic engineering construction work is carried out in large numbers,and road inspection,maintenance,and management work have become more demanding.Asphalt pavement segregation detection is an important part of ensuring pavement performance.Research on rapid extraction of accurate pavement texture and structural information is of great significance for non-destructive,rapid and accurate detection of pavement segregation.Focusing on the theme of segregation determination,with the starting point of improving detection efficiency,the use of digital images to study pavement image preprocessing,structure information acquisition methods and asphalt pavement segregation discrimination methods.Image preprocessing plays a role in noise reduction and quality enhancement,and is a prerequisite for subsequent research;the acquisition of pavement structure information is an important part of segregation determination.Under the general trend of rapid non-destructive testing,the use of digital images helps to quickly acquire new road surfaces Construction information and judgment discrimination;in terms of road surface separation judgment,multi-scale edge detection and texture image entropy are introduced to improve the road surface image separation judgment level.For newly built asphalt pavement,due to less pollutants,the asphalt color is relatively fresh,and the grayscale information can reflect the image brightness and darkness caused by the difference in the depth of the pavement structure.Therefore,a pavement structure based on the image gray matrix and neural network Depth estimation method.Using the BP neural network,the input and output samples are obtained by using the oblique photography real scene modeling and gray-scale image,and a learning and training model is established.By inputting the gray-scale matrix of the image,the road structure depth information is obtained by training.On this basis,combining the estimated structural depth and the coefficient of variation determination,through a large number of experimental analysis,a new isolation criterion is proposed.Experiments show that this method can achieve the purpose of quickly and accurately judging the segregation of newly-built asphalt pavement.For the old asphalt pavement,due to long-term use,the pavement is contaminated with stains,wheel tracks,etc.,the grayscale information in the image is not enough to distinguish the pavement texture.In view of this,the edge detection Lo G operator is used to identify the image texture information features.By using different Operator size,get image texture information.Comparing the edge detection images of different sizes,it is found that the phenomenon of partial separation and analysis between two edge detection images of the same sample is more serious.According to this feature,the edge detection texture image and image entropy are combined to obtain the texture image entropy value.Asphalt segregation status,through experimental analysis,this method can more reliably evaluate the segregation status of the old asphalt pavement,and provide a basis for the refined asphalt pavement usage evaluation and maintenance.