Abstract:Due to being affected by factors such as imaging distance, lighting, ground features, and environment, objects in the same class may have certain differences, and different classes of objects often produce similar visual features in remote sensing images. This phenomenon leads to an uncertainty problem in segmentation of remote sensing images, i.e., intra-class heterogeneity and inter-class blurring. To alleviate this problem, a fuzzy multiscale convolution neural network (FMCNet) is proposed in this paper. By extracting receptive fields of different scales, sizes and aspect ratios, the detailed information in remote sensing objects is fully represented. The relationship between their adjacent pixels is effectively expressed by fuzzy logic learning to alleviate the uncertain segmentation. The proposed method achieves overall accuracies of 85.33%, 86.31%, and 85.39% on the Vaihingen, Potsdam, and Gaofen Image datasets respectively. It demonstrates superior performance compared to existing popular methods. Influencés par des facteurs tels que la distance d'imagerie, l'éclairage, les caractéristiques du sol et l'environnement, les objets de la même classe peuvent présenter certaines différences, et différentes classes d'objets produisent souvent des caractéristiques visuelles similaires dans les images de télédétection. Ce phénomène entraîne le problème de l'incertitude dans la segmentation des images de télédétection, notamment en ce qui concerne l'hétérogénéité intra-classe et le flou inter-classe. Pour atténuer ce problème, nous proposons dans cet article un réseau de neurones convolutifs flous multiscales (FMCNet). En extrayant le champ récepteur de différentes échelles, tailles et ratios d'aspect, les informations détaillées des objets de télédétection sont pleinement représentées. La relation entre leurs pixels adjacents est efficacement exprimée par l'apprentissage de la logique floue pour atténuer la segmentation incertaine. La méthode proposée atteint des précisions globales de 85,33%, 86,31% et 85,39% sur les ensembles de données Vaihingen, Potsdam et Gaofen Image, respectivement. Elle démontre une performance supérieure par rapport aux méthodes populaires existantes.

A multiscale bidirectional fuzzy-driven learning network for remote sensing image segmentation

Fuzzy neighbourhood neural network for high-resolution remote sensing image segmentation

Remote sensing image segmentation based on the fuzzy deep convolutional neural network

Remote Sensing Image Segmentation Network Based on Adaptive Multiscale and Contour Gradient

A Semantic Segmentation Approach Based On Deeplab Network In High-Resolution Remote Sensing Images

Remote Sensing Image Semantic Segmentation Method Based on a Deep Convolutional Neural Network and Multiscale Feature Fusion

Densely Based Multi-Scale and Multi-Modal Fully Convolutional Networks for High-Resolution Remote-Sensing Image Semantic Segmentation

Deep Bilateral Filtering Network for Point-Supervised Semantic Segmentation in Remote Sensing Images

Scale-Aware Neural Network for Semantic Segmentation of Multi-Resolution Remote Sensing Images

MFCSNet: Multi-Scale Deep Features Fusion and Cost-Sensitive Loss Function Based Segmentation Network for Remote Sensing Images

FMCNet: A Fuzzy Multiscale Convolution Network for Remote Sensing Image Segmentation

Semantic Segmentation of Remote-Sensing Images Based on Multiscale Feature Fusion and Attention Refinement

Boundary-Aware Multiscale Learning Perception for Remote Sensing Image Segmentation.

A deep learning based framework for remote sensing image ground object segmentation

An Attention-Fused Network for Semantic Segmentation of Very-High-Resolution Remote Sensing Imagery

Dual-Path Multi-Level Feature Fusion Network for Semantic Segmentation of Remote Sensing Image

Multi-scale Spatial Aggregation Network for Remote Sensing Image Segmentation

Superpixel-enhanced Deep Neural Forest for Remote Sensing Image Semantic Segmentation

Encoder- and Decoder-Based Networks Using Multiscale Feature Fusion and Nonlocal Block for Remote Sensing Image Semantic Segmentation

A Multi-Scale Deep Feature Learning and Semantic Enhancement Approach for Remote Sensing Scene Classification