基于毫米波雷达获取料堆DEM的插值方法研究

针对现有的全自动堆取料机技术的发展过程中存在料堆高程模型获取困难、测量装置环境适应性差、成本高和模型精度低的问题,提出利用77GHz毫米波雷达、差分北斗和角度编码器的集成技术获取大型料堆表面的离散点云数据。推导出多传感数据融合获取离散点云数据的公式,并提出采用量子化鸽群优化Kriging插值算法获取料堆的数字高程模型(DEM),采用交叉验证的方式对比分析了优选参数后改进的普通Kriging、普通Kriging、反距离加权、基于三角剖分的线性和自然邻域插值算法,改进后的Kriging插值算法均方根误差低于0.37m,均方误差低于0.14m,均方根误差相比普通Kriging插值算法降低了39.9%。在现场的测试过程中,该方法可不受天气和粉尘的影响得到料堆的DEM,能够满足堆取料机全自动项目对精度的需求。

Research on Interpolation Method for Obtaining DEM of Material Pile Based on Millimeter Wave Radar

Aiming at the problems of difficulty in obtaining stack height model, poor environmental adaptability of measuring device, high cost and low model accuracy in the development process of the existing automatic stacker-reclaimer technology, the use of 77GHz millimeter wave radar, differential BeiDou and angle coding is proposed. The integrated technology of the device obtains discrete point cloud data on the surface of a large stockpile. Derive a formula for solving multi-sensor data fusion to obtain discrete point cloud data. It is proposed to use the hybrid quantized pigeon-inspired optimization algorithm to optimize the Kriging interpolation algorithm to obtain the digital elevation model(DEM)of the stockpile. The method of cross-validation is used to compare and analyze the improved ordinary Kriging, ordinary Kriging, inverse distance weighting, linear and natural neighborhood interpolation algorithms based on triangulation after selecting the parameters. The root mean square error of the improved Kriging interpolation algorithm is less than 0.37m, the mean square error is less than 0.13m, and the root mean square error is 39.9% lower than the ordinary Kriging interpolation algorithm. During the on-site test, the proposed method can obtain the DEM of the pile without the influence of weather and dust, and it is meets the requirements of accuracy in the project of automatic pile-reclaimer.