Research Of3-D Reconstruction Algorithm Based On Materials Stack Measurement System

In the production process of logistics information management, effective stockpile measurement for large enterprises in the logistics information management is a very important part. In order to improve the automation standard for materials management, how to accurately measure the material stack’volume become a pressing issue.After analysis a variety of existing three-dimensional measurement methods, the subject use a material stack volume measurement based on computer vision technology. The program is that:the use of artificial light through the black and white grid surface exposure to the material stack, man-made features point, and then read the image using the camera around. These feature points by finding and matching, to obtain three-dimensional coordinates of points, and then three-dimensional reconstruction of the material stack. On this basis, design a volume measuring device, and use this device to complete the three-dimensional scattered point data collection.The data from this device can’t be directly calculated, so these data need to be pretreated. In the experiment we can’t guarantee that the stage completely flat plane parallel with the material stack, resulting in scattered point data is deflection. So the project calculating the tilt angle with the least square method by the boundary points. Then complete the correction of the entire material stack. At the same time, choosing wavelet filtering algorithm for the noise on the material heap, to prepare for the next step.When the volume of material stack is calculated and three-dimensional reconstruction is taken, this paper presents a volume calculation based on interpolation methods. In the calculation process, the interpolation algorithm of choice is very important. Interpolation algorithm directly affects the accuracy of volume calculation accuracy. In this paper, a simple algorithm based on inverse distance method and the statistical Kriging and RBF network interpolation is selected. To ensure the feasibility and accuracy of the algorithm, the paper design artificial data validation and analysis of algorithms. When the algorithm is theoretically possible, using the algorithm on the actual material stack. After calculated, the algorithm’s accuracy is meeting to the actual requirements.