Remote Sensing Interpretation Of Multi-Source Data For Agricultural Land Survey:a Method Based On WebGIS

Visual interpretation is the basic method of land use classification based on large-scale remote sensing data, which has a problem of efficiency and accuracy for a long time. One important reason is that interpretation process is limited by the traditional paper-based operating platform, and unable to effectively integrate multi-source heterogeneous data to support complex logic. WebGIS can make up for deficiencies in this regard, namely through the space integration of online/offline heterologous and heterogeneous data, as well as a combination of multi-temporal remote sensing images to collect more visual evidence, which maximum extent minimizes land interpretation uncertainty, and integrates the results of interpretation into the database instantly.In this paper, we regard large-scale remote sensing survey of agricultural land as a starting point, study spatial integration methods and techniques of multi-source heterogeneous data based on WebGIS platform, explore and build a logical inference rule of multivariate compound evidence for visual interpretation of land classification to improve the number of categories and classification accuracy. The scientific issues of this paper is how to build credibility in its category semantic by more evidence of composite reasoning for unrecognized feature on the spectrum and morphology. This involves knowledge using of agricultural culture and practice characteristics in the study area, and not simply visual judgment. The findings will refine the interpretation rules used in the expert system/knowledge base for remote sensing interpretation.Firstly, this paper elaborate the basic theory related to land use investigation and visual interpretation of remote sensing images, and outlines the applications of "3S" technology (geographical information system, remote sensing and global positioning system) in land use survey.Secondly, from the practical application needs, this study designs the online collection system for agricultural land data. Based on the theory of visual interpretation of remote sensing images, combined with the characteristics of the WebGIS platform, this study puts forward a new method on agricultural land survey through the integration of multi-source auxiliary information, which is the methodological basis of this study.Finally, the study explores the above interpretation methods through the agricultural land survey of Shanghai. On the basis of on-line collecting of agricultural land data, the study completes the accuracy verification, the result has a high data quality and the rationality of the research method is proved.In addition, by participating in the indoor interpretation and field investigation, the study improve the interpretation rules of agricultural land in Shanghai based on aerial images, and provide a reference for automatic interpretation algorithm designing in the late.