Fusion Classification of HSI and MSI Using a Spatial-Spectral Vision Transformer for Wetland Biodiversity Estimation

Yunhao Gao,Wei Li,Xiukai Song,Yinyin Feng,Jianbu Wang,Jianlong He,Xiangyang Jiang

doi:10.3390/rs14040850

Abstract

The rapid development of remote sensing technology provides wealthy data for earth observation. Land-cover mapping indirectly achieves biodiversity estimation at a coarse scale. Therefore, accurate land-cover mapping is the precondition of biodiversity estimation. However, the environment of the wetlands is complex, and the vegetation is mixed and patchy, so the land-cover recognition based on remote sensing is full of challenges. This paper constructs a systematic framework for multisource remote sensing image processing. Firstly, the hyperspectral image (HSI) and multispectral image (MSI) are fused by the CNN-based method to obtain the fused image with high spatial-spectral resolution. Secondly, considering the sequentiality of spatial distribution and spectral response, the spatial-spectral vision transformer (SSViT) is designed to extract sequential relationships from the fused images. After that, an external attention module is utilized for feature integration, and then the pixel-wise prediction is achieved for land-cover mapping. Finally, land-cover mapping and benthos data at the sites are analyzed consistently to reveal the distribution rule of benthos. Experiments on ZiYuan1-02D data of the Yellow River estuary wetland are conducted to demonstrate the effectiveness of the proposed framework compared with several related methods.

Highlights

Coastal wetland is a transitional area between terrestrial and marine ecosystems, which has a complex environment and monitoring elements [1]
The land-cover mapping obtained by support vector machine (SVM), LBP-extreme learning machine (ELM), and S2FL tends to be rather noisy, resulting in serious landscape fragmentation
Regarding the land-cover mapping produced by Two-branch convolutional neural network (CNN) and depth feature interaction network (DFINet), the fragmentation and artifacts are alleviated

Summary

Introduction

Coastal wetland is a transitional area between terrestrial and marine ecosystems, which has a complex environment and monitoring elements [1]. Accurate biodiversity monitoring of coastal wetlands is of great significance in water conservation [2], biodiversity conservation [3], and blue carbon sink development [4]. The traditional on-site monitoring receives data by stations and sections, which is time-consuming and laborious. Remote sensing technology has the advantages of large-area coverage, spatio-temporal synchronization, and high spatial-spectral resolution [5], providing highly relevant information for a wide range of wetland monitoring applications. Biodiversity estimation based on remote sensing achieves economic and real-time data collection. A lot of works have been developed for biodiversity estimation based on remote sensing

Results

Discussion

Conclusion