Development And Experimental Study On The Manufactured Sandsize And Shape Detection System

Asphalt mixture is the most commonly used road construction materials in the world.40%~50% volume of asphalt mixture is fine aggregate.However,the natural sand source reduces dramatically as the huge market demand.Therefore,it is an inevitable trend to replace natural sand by manufactured sand.However,the qualities of manufactured sand’s size and shape vary greatly.Testing manufacured sand is the key to ensure the quality of asphalt mixture.Conventional vibration screening method can only achieve the manufacured sand particle size distribution detection.By contrast,image method can simultaneously achieve particle size and shape detection.In this dissertation,image detection algorithms aimed to detect the manufactured sand with diameter between 0.6~4.75 mm are researched,and the corresponding detection system is developed.A manufactured sand vibration dispersion system is designed,and a CCD camera is used to collect falling sand images.A less shadow backlight system is designed to reduce the influence of the sand surface color on the detection results,the the manufactured sand size and grain detection hardware is developed.In software part,Gaussian filtering is used to denoise the gray image.Also,Otsu method is used to segment the filtered image and get the binary image.The particle geometrical characteristic is used to discriminate the incomplete particles at the image boundary then remove it.Hu moments are used to identify and eliminate the particles that are repeatedly photographed in the adjacent picture.Thus,the adhered particles are separated by finding the convex hull and pits of the particles.A new image calibration method has been developed.Based on Visual Studio C + +,Open CV library and Qt,a manufacured sand size and shape detection system is developed.In addition,the repeatability test and contrastive precision test with the QICPIC(Sympatec Gmb H)of 0.6~4.75 mm single-grading and graded particle has been completed in this project.The experimental results show that the maximum repeatability errors of particle size detection are 3.46% and 0.51% for single-grading and graded particle,and the maximum repeatability errors of grain shape detection are 2.97% and 0.85%.The repeatability increases as the shape roundness goes up.Compared with the particle size result of QICPIC,the maximum deviations of single-grading and graded particle are 7.19% and 6.02%,and the maximum deviations of grain shape are 3.08% and 2.42% respectively.Aiming at the difference between the image method and the sieving method,a kind of particle size correction method has been proposed.This method satisfies the accuracy demand of the particle size detection.Then flowing time method which measures fine aggregate angularity in Test Methods of Aggregate for Highway Engineering(JTG E42-2005)is used as a reference to analyze correlation between different grain shape characterization parameters and flowing time.The result shows that ratio of equivalent ellipse axes is the optimal parameter.In one word,this developed detection system can meet the lab requirement of manufactured sand size and shape detection,and effectively monitor the quality of sand.