Identifying Urban Wetlands through Remote Sensing Scene Classification Using Deep Learning: A Case Study of Shenzhen, China

Abstract

Urban wetlands provide cities with unique and valuable ecosystem services but are under great degradation pressure. Correctly identifying urban wetlands from remote sensing images is fundamental for developing appropriate management and protection plans. To overcome the semantic limitations of traditional pixel-level urban wetland classification techniques, we proposed an urban wetland identification framework based on an advanced scene-level classification scheme. First, the Sentinel-2 high-resolution multispectral image of Shenzhen was segmented into 320 m × 320 m square patches to generate sample datasets for classification. Next, twelve typical convolutional neural network (CNN) models were transformed for the comparison experiments. Finally, the model with the best performance was used to classify the wetland scenes in Shenzhen, and pattern and composition analyses were also implemented in the classification results. We found that the DenseNet121 model performed best in classifying urban wetland scenes, with overall accuracy (OA) and kappa values reaching 0.89 and 0.86, respectively. The analysis results revealed that the wetland scene in Shenzhen is generally balanced in the east–west direction. Among the wetland scenes, coastal open waters accounted for a relatively high proportion and showed an obvious southward pattern. The remaining swamp, marsh, tidal flat, and pond areas were scattered, accounting for only 4.64% of the total area of Shenzhen. For scattered and dynamic urban wetlands, we are the first to achieve scene-level classification with satisfactory results, thus providing a clearer and easier-to-understand reference for management and protection, which is of great significance for promoting harmony between humanity and ecosystems in cities.

Large Kernel Separable Mixed ConvNet for Remote Sensing Scene Classification

“…As discussed in Section II-C, the size of the large kernel convolution is a critical parameter that governs the receptive field and the efficiency of feature extraction within the corresponding convolution operation. To scrutinize the effect of varying K on classification performance, we set K to [3,5,7,9,11] and perform pertinent experiments across three datasets. The graphs in Fig.…”

Hands-off, artificial construction, or penalty? How to deal with the increasingly polluted coastal wetland ecosystem in China

“…Zhang et al (2022) used high-resolution numerical models to simulate wetland fluxes. Yang et al (2022) proposed an urban wetland identification framework based on an advanced scene-level classification scheme to identify wetlands. These studies mainly evaluated wetland ecosystems from the aspects of geospatial technology, isotope tracer technology, numerical simulation, and so on.…”

Esearch on the optimization path of campus football teaching based on deep learning mode

“…In other words, by layer-by-layer feature transformation, the feature representation of a sample in the original space is transformed into a new feature space, thus making classification or prediction easier. Finally, deep learning can achieve our automation requirements for complex transaction processing classification by designing and establishing the right amount of neuronal computation nodes and multi-layer operational hierarchy, selecting the right input and output layers, and establishing a functional relationship from input to output through learning and tuning of the network [23].…”