Compaction Detection Research Based On Digital Image Processing

Since the 21st century,China’s road transportation and construction industry has developed rapidly,and the road mileage has been greatly expanded.High-grade highway construction requires high quality,and the quality of roadbed compaction has a great impact on the quality of highway construction.The traditional subgrade compaction detection methods have shortcomings such as time-consuming,poor representativeness and destructiveness,which cannot meet the needs of current highway development.With the rapid development of digital image processing technology,it has become a non-contact and high-accuracy optical measurement method,which can provide convenience for the quantitative research of civil engineering.In this paper,the Guanglian Expressway granite residual soil is taken as the research object,the characteristic geometric parameters of the compacted soil sample image are extracted by digital image processing,and the precise mathematical relationship between the image parameters and the dry density and its discriminant model are established through statistical analysis,so as to realize the Non-destructive testing of compaction.Granite residual soil has many pores,complex mineral composition,and fuzzy boundaries between pores and minerals.It is difficult to accurately extract effective parameter information related to dry density by conventional digital image methods.In this paper,through the analysis of the porosity change law of the granite residual soil sample during the compaction process,it is concluded that the void pixel area Se and the void ratio E in the image are the two discriminant parameters that are most closely related to the dry density change.At the same time,the digital image processing scheme is studied to accurately extract the parameter information,that is,after gray-scale conversion of the original image,the histogram equalization is used to enhance the image contrast,and then the local adaptive threshold method is used for binarization.The foreground part is used for contour recognition and area calculation,which can realize quantitative analysis of image parameters,and solve the problems of uneven illumination and large gap between light and dark in the image.In order to overcome the ambiguity of the pore boundary of the image and the difficulty of collecting effective image data,the research proposes a method of preparing a transparent test mold and using a fixed-distance and fixed-orientation elevation image acquisition method in the indoor test to ensure the high analysis accuracy of the test.original image base.Statistical analysis of a large number of image samples confirmed that the two discriminant parameters had a good linear relationship with dry density,and a two-parameter dry density discriminant model was established accordingly.The inspection test confirmed that when the error is less than or equal to 0.01g/cm3,the model discrimination accuracy rate exceeds 80%,and the dry density discrimination of the sample under indoor conditions has been successfully realized,which provides a theoretical basis for the non-destructive testing of compaction degree using the digital image method.Base.The soil conditions at the engineering site are more complex,and a new research plan is established with reference to indoor research.It is established that Se and soil dry density are in a power function relationship,and a power function dry density discrimination model is established accordingly.Relying on the high embankment project of Huadu to Conghua Section of Guangzhou to Lianzhou Expressway,the compaction degree detection plan of the construction site is proposed and tested on the spot.The research shows that the more samples analyzed,the higher the discrimination accuracy.When the dry density error is less than or equal to 0.01g/cm3 and when the compaction degree error is less than or equal to 1%,the discrimination accuracy of dry density and compaction degree of a single measuring point is not high;the discrimination accuracy of dry density and compaction degree of the area represented by the average value of 10 measuring points is 100%.