Abstract: Interpretation of natural resources features is the most basic, most widely used, most involved, and most technically difficult task in survey and monitoring work. It faces many challenges such as application understanding, accurate interpretation, and fine interpretation. Starting from the application scenarios of natural resource survey and monitoring, firstly, this article elaborates its research progress from three aspects:scene-features coupling interpretation mode, semantic understanding expression cognition, and human-computer collaborative interpretation methods. Then, it discusses six key research directions and basic ideas, including scene-feature coupling interpretation, knowledge-driven semantic understanding, human-machine collaborative intelligent interpretation, indoor and outdoor integrated and 3D environment interpretation, quantitative inversion of key parameters, high-confidence quality control and authenticity verification. Finally, summaries are given. This paper aims to provide some ideas for the research on intelligent interpretation of natural resources survey and monitoring features, and promote the solution of bottlenecks that plague the application of remote sensing image automatic interpretation methods in natural resource survey and monitoring applications.

Key words: natural resources survey and monitoring, intelligent interpretation, scene-features coupling interpretation, semantic understanding, human-computer collaboration, indoor and outdoor integrated interpretation, parameter inversion, authenticity verification