宇航学报

宇航学报 ›› 2019, Vol. 40 ›› Issue (11): 1358-1366.doi: 10.3873/j.issn.1000-1328.2019.11.012

• 制导、导航、控制与电子 • 上一篇    下一篇

卫星遥感农业监测系统中实例检索算法研究

臧晶，李成华，田野   

  1. 1. 沈阳农业大学工程学院，沈阳 110866；2. 沈阳理工大学信息科学与工程学院，沈阳 110159；
    3. 沈阳理工大学机械工程学院，沈阳 110159
  • 收稿日期:2019-06-06 修回日期:2019-08-27 出版日期:2019-11-15 发布日期:2019-11-25
  • 基金资助:
    国家自然科学基金资助项目（51075282,61471247, 61501307）;辽宁特聘教授项目;辽宁企业项目博士后资助;沈阳博士后基金资助;辽宁省一般项目（L201727）;辽宁省教育厅高等学校基本科研项目（LG201708）;辽宁省自然科学基金指导计划项目（20180550520）

Research on Case Retrieval Algorithm in Satellite Remote Sensing Monitoring System for Agriculture

ZANG Jing, LI Cheng hua, TIAN Ye   

  1. 1. College of Engineering, Shenyang Agricultural University, Shenyang 110866, China; 2. School of Information Science and
    Engineering, Shenyang Ligong University, Shenyang 110159, China; 3. School of Mechanical Engineering,
    Shenyang Ligong University, Shenyang 110159, China
  • Received:2019-06-06 Revised:2019-08-27 Online:2019-11-15 Published:2019-11-25

PDF

164

摘要： 针对智慧农业中需求获取手段、表达方法等导致智能设计算法复杂度高、可用性低等问题，将卫星遥感技术和网络技术结合实现农作物及其生长环境参数的实时监测，以优化播种机设计的需求特征，并以此为基础，提出基于混合相似度的实例检索模型。在该模型中，设计了多级实例库结构，并以特征间距离相似度、特征间可拓关联度、特征间可拓距测度为组成要素，构造多级混合相似度的层次结构与计算方法，进行多级实例检索。将其应用到铲式播种机的数字化设计系统中，可获得与目标实例最相似的实例。实验结果表明，在实例的特征值数据类型多样化的情况下，此算法能够有效地检索出合适的历史实例，提高检索结果的可用性。

关键词: 遥感, 混合相似性, 实例检索, 基于实例推理的设计

Abstract: The acquisition means and expression methods of demands in intelligent agriculture result in the high complexity and low usability of the intelligent mechanic design methods. To reduce this complexity, improve the usability, and optimize the demand features of planter design, the real-time monitoring of the crops and the parameters in their growth environment is employed by combining the satellite remote-sensing technology with the network technology, and a hybrid similarity retrieval model is proposed in this paper. Moreover, to perform the multi-hierarchical case retrieval, the hierarchy and calculation methods of the algorithm are constructed in this model by integrating the distance similarity between features, the extension correlation degree between features, and the extension distance between features based on the design of the multi-hierarchy case base. The proposed method is applied to the digital design system of a spade planter and the cases most similar to the target case are obtained. The experimental results indicate that under the condition of the diverse data types of cases feature values, the proper history cases are obtained and the usability of the retrieval results is improved.

Key words:  Remote sensing, Hybrid similarity, Case retrieval, Case-based design (CBD)

中图分类号: