基于双目视觉的不规则堆料体积测量研究

Research on Volume Measurement of Irregular Pile Based on Binocular Vision

摘要:

针对不规则堆料的体积难以直接测量的问题,提出一种基于双目视觉的不规则堆料体积测量方法。首先,使用双目相机采集堆料的俯视图像,对图像进行预处理；接着通过立体匹配求出堆料图像的视差图,并采用中值滤波和最邻近插值算法对于视差图中的空洞进行填充,同时计算出视差图对应的三维点云坐标；再采用改进的OTSU阈值分割算法将视差图中的堆料区域提取出来,计算出堆料区域的平均高度和面积,最后计算出堆料的体积。实验结果表明,体积测量误差小于5%,可以快速有效地测量出堆料的体积,能够满足工业检测领域的需求,同时为准确快速地测量堆料体积提供了一种新的思路。

Abstract:

Aiming at the problem that the volume of irregular piles is difficult to measure directly, a method for measuring the volume of irregular piles based on binocular vision is proposed. Firstly, a binocular camera is used to collect the top view images of the pile, and the images are preprocessed. Secondly, the disparity map of the pile image is obtained by stereo matching, and the holes of the disparity map are filled by the median filtering and the nearest neighbor interpolation algorithm, and the three-dimensional point cloud coordinates corresponding to the disparity map are calculated. Thirdly, the improved OTSU threshold segmentation algorithm is used to extract the pile area in the disparity map, the average height and area of the pile are calculated. Finally, the volume of the pile is calculated. The experimental results show that the volume measurement error is less than 5%, and the volume of the pile can be measured quickly and effectively based on disparity map, which can meet the needs of the industrial inspection field, and provides a new idea for accurately and quickly measuring the volume of the pile.