| Analyzing the existing volume measurement methods of large material stacks, this paper researches the volume measuring method of large and complex shape material stacks based on tri-vision. First, this project designs an experimental device. In order to solve the problem of large-scale material stack surface with slower transformed gray value and fewer feature points, this paper produces artificial feature points through the method of black and white grid projection. This paper studies the camera calibration technique that has considered the camera radial distortion and strikes the camera’s internal and external parameters by Tsai two-step calibration method and particle swarm optimization method. This method is simple and the requirement on the calibration object is not strict, so it is suitable for large stockpile volume calculation.Collecting the image information from three different directions, the project preprocesses the images in order to enhance the image feature point information. This paper selects the artificial registration method to extract the matching points. In two approximate conditions, it improves edge matching algorithm guided by points and corresponding points matching algorithm tied by edge to increase in the matching angle points using extracted matching points. And the project analyzes the influence on the stockpile volume errors reduced by the approximation of two conditions. Then, according to calibrated camera internal and external parameters, the project restores matching point pixel values into three-dimensional coordinates.Finally, according to the restored three-dimensional scatter point coordinates, this paper presents a volume calculation and three-dimensional reconstruction based on interpolation method according to the restored three-dimensional scatter point coordinates. In the process of calculation, the accuracy of interpolation algorithm directly affects the accuracy of volume calculation. This article uses precise radial basis function interpolation method. While using the algorithm on the simulated material stack, the volume measurement error is within 10%, so the algorithm’s accuracy can meet the actual requirements.To some extent, this research realizes the automatic measurement of large and complex shape material stacks. Modeling and testing in the laboratory to verify the feasibility and accuracy of this program. The method is simple, low cost. |
