数据融合及其在农情遥感监测中的应用与展望

Data fusion and its application prospect in agricultural condition monitoring using romote sensing

长期以来,由于对数据融合一直没有一个严格的统一的定义,对于数据融合的理解、表达存在各种差异,在一定程度上影响了学术交流的顺利开展和应用渠道的畅通。另一方面,数据融合在遥感领域的长期发展已取得了丰硕的成果,将这些成果应用于农情遥感监测中,是非常有意义的。该文详细介绍了数据融合的概念及其发展过程,包括数据融合的定义、数据融合层次和融合方法;分析了当前数据融合在遥感尤其农情遥感监测中的应用现状,并针对农情遥感监测的特点,展望数据融合在农情遥感监测中的应用前景。

The concept of data fusion was proposed in 1970s and developed in 1990s. Many researches have been done in this area. However, for a long time there is not a specific definition of data fusion that is commonly accepted and used by academic societies of information technology. The description of data fusion is usually different from one to another, which confuses the communication in the scientific community and influences the technology transfer to industry. On the other hand, how to apply the data fusion technology achievements to remote sensing in agricultural condition monitoring is also a very important area. In this paper, the concept of data fusion and its development are reviewed in detail, including the definition, fusion levels, and methods. Then the paper analyses the status of its application in agricultural condition monitoring: high spatial and spectral resolution images begin to be used in agricultural condition monitoring together with low resolution ones, and some ancillary data play a more important role in helping with crop identification and image classification. For how to better use data fusion in agricultural condition monitoring, which is not only operational, on large scale, with great complexity, but also requires high accuracy, the application prospects are discussed by the authors.