Research on Probability Identification Method of Aggregate Accumulation Gradation Based on Machine Vision
The traditional sampling and sieving method is difficult to measure the aggregate gradation in the storage yard quickly and accurately.Therefore,this paper establishes a binocular stereo vision acquisition system,which projects points in the real three-dimensional space onto the two-dimensional plane of the camera and performs stereo matching to obtain a disparity map.Combined with binocular stereo vision distance measurement,the three-dimensional model of the accumulation body is obtained.Then,the conical accumulation body is photographed from the above to capture images,and the aggregate image segmentation is carried out by using the deep learning method.The proportion of the area of the particle aggregate in the image is counted,and the multivariate linear regression model is obtained by multivariate linear fitting of the coarse particle concentration in different height ranges of the same belt.Compared with the results of manual sieving measurement,it shows that the method in this paper has feasibility and application prospect.
NETL
NSTL
万方数据
