基于机器视觉的散装粮随机扦样方法研究

针对粮食收购过程中扦样设区选点不科学、不合理导致的扦取样品代表性不足和存在人为舞弊风险的问题，提出一种结合双目视觉与图像分割技术的散装粮随机扦样方法。首先使用双目相机获取装粮区域图像信息并校正，利用Unet网络模型实现校正后左图像目标分割，再使用Opencv计算目标区域4个角点像素坐标，根据扦样规则将目标区域划分为多个扦样区域并随机生成扦样点，最后针对BM匹配算法生成视差图效果较差的问题，采用SGBM(Semi-Global Matching)半全局立体匹配算法对校正后左右图像立体匹配，根据匹配结果完成扦样点三维空间定位。实验结果表明，所述方法针对装粮区域有较好的识别效果，并且实现了在3 m范围内扦样点的随机选取与定位，对粮食扦样环节的自动化和智能化发展提供了技术支撑。

Research about random sampling method of bulk grain based on machine vision

In order to solve the problem of inadequate representation of samples and the risk of fraud caused by unscientific and unreasonable location selection in grain purchasing process, a random sampling method for bulk grain was proposed combining binocular vision and image segmentation technology. Firstly, binocular camera was used to obtain image information of grain loading area and correction, and the Unet network model was used to achieve target segmentation of corrected left image. Then, the Opencv was used to calculate pixel coordinates of four corner points of the target area, and the target area was divided into multiple sampling areas according to the sampling rules and randomly generating the sampling points. Finally, aiming at the problem that the disparity map generated by BM Matching algorithm was poor, the semi-global stereo Matching algorithm of SGBM(Semi-global Matching) was used to match the stereo of the corrected left and right images, and the 3D space positioning of the sample points is completed according to the Matching results. The experimental results show that the method has a good identification effect of grain loading area, and randomly selection and positioning of sampling points within 3m were achieved, which provides a technical support for the automation and intelligent development of grain sampling.