Abstract:In recent years,semantic segmentation technology plays an important role in land resource management tasks. However, many classic semantic segmentation models often fail to obtain satisfactory results for remote sensing images with large difference of object scale span. The statistical texture involved multi-granularity attention network has been proposed to improve this situation. Statistical texture involved multi-granularity attention network has an encoder-decoder structure which is similar to DeeplabV3+ 2018. On this basis,texture excitation module and multi-granularity attention module are proposed to improve the accuracy of semantic segmentation. Texture excitation module is designed to extract statistical texture feature that are hidden in the image so that they can be combined with spatial structure feature for better prediction. Specifically, this module calculates the pixel co-occurrence matrix in the horizontal, vertical, diagonal, and anti diagonal directions. Multi-granularity attention module provides a more comprehensive understanding of the image by simultaneously focusing on the image at coarse,medium and fine granularity. Specifically, this module designs coarse granularity, medium granularity, and fine granularity spatial attention mechanisms and coarse granularity, medium granularity, and fine granularity channel attention mechanisms based on the core idea of self attention mechanism. Statistical texture involved multi-granularity attention network is compared with several of the most advanced deep learning methods on the Vaihingen data set and Potsdam data set.Compared with the best performing method, our method has increased by 0.73%, 1.88% and 0.013 in pixel accuracy,mean intersection over union and F1 score respectively on Vaihingen data set; our method improves the pixel accuracy from 87.82% to 88.50%, the mean intersection over union from 74.53% to 75.82%, and the F1 score from 0.8488 to 0.8573 on Potsdam data set.

Orientation Attention Network for Semantic Segmentation of Remote Sensing Images?

AANet: Adaptive Attention Networks for Semantic Segmentation of High-Resolution Remote Sensing Imagery

AANet: an Attention-Based Alignment Semantic Segmentation Network for High Spatial Resolution Remote Sensing Images

MOA-Net: Multilevel Object Aware Network for Remote Sensing Image Semantic Segmentation

Semantic segmentation of remote sensing images combined with attention mechanism and feature enhancement U-Net

MASANet: Multi-Angle Self-Attention Network for Semantic Segmentation of Remote Sensing Images

Semantic Segmentation With Attention Mechanism for Remote Sensing Images

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

A position-aware attention network with progressive detailing for land use semantic segmentation of Remote Sensing images

Hybridizing Cross-Level Contextual and Attentive Representations for Remote Sensing Imagery Semantic Segmentation

DEANet: Dual Encoder with Attention Network for Semantic Segmentation of Remote Sensing Imagery

XANet: An Efficient Remote Sensing Image Segmentation Model Using Element-Wise Attention Enhancement and Multi-Scale Attention Fusion

SACANet: scene-aware class attention network for semantic segmentation of remote sensing images

Global-Local Attention Network for Semantic Segmentation in Aerial Images.

Shallow Multiplexing and Multiscale Dilation Convolution Combined Attention Based Oriented Object Detection in Remote Sensing Images

Multi-scale Spatial Aggregation Network for Remote Sensing Image Segmentation

Statistical Texture Involved Multi-Granularity Attention Network for Remote Sensing Semantic Segmentation

AFNet: Adaptive Fusion Network for Remote Sensing Image Semantic Segmentation

Dual Path Attention Net For Remote Sensing Semantic Image Segmentation

Hybrid Multiple Attention Network for Semantic Segmentation in Aerial Images

Global Semantic-Sense Aggregation Network for Salient Object Detection in Remote Sensing Images