Two Strategies for Remote Sensing Classification Accuracy Improvement of Salt Marsh Vegetation: A Case Study in Chongming Dongtan

Remote sensing technology has become the primary tool for salt marsh vegetation classification at large scales. However, there is still a major problem in differentiating between different spectra for the same vegetation and the same spectrum for different vegetation, when classifying salt marsh vegetation in remotely sensed images. In this paper, two strategies for this problem were proposed. One was through the integration and application of multi-seasonal images based on decision tree method, and another was through the integration of auxiliary data with remote sensing based on fuzzy mathematical theory. It was proved that the two strategies can improve the classification accuracy of salt marsh vegetation to some extent and have a good popularization value.