基于DBSCAN的农机作业轨迹聚类研究

发布时间：2018-05-05 01:39

  本文选题：农业机械 + 作业轨迹　； 参考：《农机化研究》2017年04期

【摘要】：农业机械在田间作业过程中,时间和空间维度上产生大量的作业数据,对农业机械作业轨迹数据进行聚类分析在农机作业状态分析和效率研究中具有重要意义。为此,应用DBSCAN(Density-Based Spatial Clustering of Applications with Noise)算法对模拟农业机械作业轨迹进行分析,设计了基于密度聚类的农机作业状态分类算法。对模拟数据的聚类结果表明:该方法正确分类农机作业班次内的有效作业轨迹、空行转移轨迹和停歇轨迹的精度达到98.33%、70%和100%。聚类作业轨迹反映的农机利用率为95.35%,为农机田间作业轨迹研究提供了依据。
[Abstract]:A large number of operational data are produced in the time and space dimensions of agricultural machinery in the process of field operation. Cluster analysis of agricultural machinery track data is of great significance in the analysis of agricultural machinery operation state and efficiency. Therefore, the DBSCAN(Density-Based Spatial Clustering of Applications with Noise algorithm is used to analyze the track of simulated agricultural machinery operation, and the classification algorithm of agricultural machinery operation status based on density clustering is designed. The clustering results of the simulated data show that this method correctly classifies the effective working trajectories in the shifts of agricultural machinery operations, and the accuracy of the empty transfer trajectory and the rest track reaches 98.3370% and 100% respectively. The utilization ratio of farm machinery reflected by cluster operation track was 95.35, which provided the basis for the research of farm machinery track.
【作者单位】： 新疆农业大学机械交通学院;
【基金】：国家自然科学基金项目(51465057)
【分类号】：S22;TP311.13
，

本文编号：1845625