Abstract:Accurate remote sensing (RS) monitoring of wetland ground objects is an enormous importance for ecological preservation. Wetland classification on multi-source remote sensing images (MS-RSI) includes leveraging data from different sensors for accurately describing and categorizing wetland regions. This method normally incorporates data from infrared, radar, and optical sensors to take a wide-ranging view of wetland features. Advanced image processing methodologies, comprising machine learning (ML) approaches are often implemented for analyzing these multi-source images as well as recognizing spectral and spatial patterns indicative of wetland characteristics. The interaction of various RS data increases the accuracy and robustness of wetland classification models, allowing a more complex analysis of wetland ecosystems and aiding environmental observation, conservation, and control measures. To accomplish effective training for wetland mapping through the RS, it is essential for a significant training data that comprises a numerous array of class variants. In this article, we propose an Enhanced Wetland Classification using a Deep Learning based Fusion Approach (EWC-DLFA) on MS-RSI. The proposed EWC-DLFA technique examines the MS-RSI for wetland classification using the DL model which can be used for other land cover classification types. To accomplish this, the EWC-DLFA technique utilizes the data from multiple sources such as Sentinel-1 (SAR), Landsat-8, Sentinel2 (multi-spectral), and digital elevation model (DEM). In the presented EWC-DLFA technique, a deep convolutional neural network-based EfficientNetB-5 model can be applied for the extraction of features from the multi-source images. For increasing the performance of the EfficientNet-B5 model, the marine predators algorithm (MPA) based hyper parameter tuning process can be applied. Finally, an ensemble of three ML classifiers such as extreme learning machine (ELM), multilayer perceptron (MLP), and gradient boosting machine (GBM) are used to classify the wetland into different types such as fen, bog, marsh, swamps, and upland. The performance of the EWC-DLFA technique can be validated using a large set of simulations. The resultant values pointed out that the EWC-DLFA technique reaches better performance over other models on wetland classification.

Convolutional Neural Network for Coastal Wetland Classification in Hyperspectral Image.

Multiscale 3-D-2-D Mixed CNN and Lightweight Attention-Free Transformer for Hyperspectral and LiDAR Classification

Joint Classification of Hyperspectral and Multispectral Images for Mapping Coastal Wetlands

Hyperspectral and Multispectral Classification for Coastal Wetland Using Depthwise Feature Interaction Network.

Multisource Remote Sensing Classification for Coastal Wetland Using Feature Intersecting Learning

Biscale Convolutional Self-Attention Network for Hyperspectral Coastal Wetlands Classification

Integrating Multitemporal Sentinel-1/2 Data for Coastal Land Cover Classification Using a Multibranch Convolutional Neural Network: A Case of the Yellow River Delta

Fully Dense Multiscale Fusion Network for Hyperspectral Image Classification

Multisource Feature Embedding and Interaction Fusion Network for Coastal Wetland Classification With Hyperspectral and LiDAR Data

A Hierarchical Classification Framework of Satellite Multispectral/Hyperspectral Images for Mapping Coastal Wetlands

Enhanced Wetland Classification using Deep Learning based Fusion Approach on Multi-Source Remote Sensing Images

Multi-Scale Depthwise Separable Convolutional Neural Network for Hyperspectral Image Classification

Hyperspectral Image Classification based on Multi-Scale Convolutional Features and Multi-Attention Mechanisms

A Novel 2D-3D CNN with Spectral-Spatial Multi-Scale Feature Fusion for Hyperspectral Image Classification

Hyperspectral Remote Sensing Images Deep Feature Extraction Based on Mixed Feature and Convolutional Neural Networks

Dilated Deep MPFormer Network for Hyperspectral Image Classification

Deep Multilayer Fusion Dense Network for Hyperspectral Image Classification

Object-oriented Multiscale Deep Features for Hyperspectral Image Classification

Multi-layer feature fusion for hyperspectral image classification

Classification Based On Deep Convolutional Neural Networks With Hyperspectral Image

CNN-Based Multilayer Spatial–Spectral Feature Fusion and Sample Augmentation with Local and Nonlocal Constraints for Hyperspectral Image Classification