ARTÍCULO
TITULO

A Dual-Path and Lightweight Convolutional Neural Network for High-Resolution Aerial Image Segmentation

Gang Zhang    
Tao Lei    
Yi Cui and Ping Jiang    

Resumen

Semantic segmentation on high-resolution aerial images plays a significant role in many remote sensing applications. Although the Deep Convolutional Neural Network (DCNN) has shown great performance in this task, it still faces the following two challenges: intra-class heterogeneity and inter-class homogeneity. To overcome these two problems, a novel dual-path DCNN, which contains a spatial path and an edge path, is proposed for high-resolution aerial image segmentation. The spatial path, which combines the multi-level and global context features to encode the local and global information, is used to address the intra-class heterogeneity challenge. For inter-class homogeneity problem, a Holistically-nested Edge Detection (HED)-like edge path is employed to detect the semantic boundaries for the guidance of feature learning. Furthermore, we improve the computational efficiency of the network by employing the backbone of MobileNetV2. We enhance the performance of MobileNetV2 with two modifications: (1) replacing the standard convolution in the last four Bottleneck Residual Blocks (BRBs) with atrous convolution; and (2) removing the convolution stride of 2 in the first layer of BRBs 4 and 6. Experimental results on the ISPRS Vaihingen and Potsdam 2D labeling dataset show that the proposed DCNN achieved real-time inference speed on a single GPU card with better performance, compared with the state-of-the-art baselines.

 Artículos similares

       
 
Vassilios N. Grigoriadis, Vassilios D. Andritsanos and Dimitrios A. Natsiopoulos    
Recent Digital Surface, Elevation, and Bathymetric Models (DSM/DEM/DBM) aim to provide high-resolution and accurate height and depth information needed for a variety of surveying, geodetic, geophysical, and other applications. In this study, first we aim... ver más

 
Yizhi Wang, Jia Liu, Lin Xu, Fuliang Yu and Shanjun Zhang    
Streamflow modelling is one of the most important elements for the management of water resources and flood control in the context of future climate change. With the advancement of numerical weather prediction and modern detection technologies, more and m... ver más
Revista: Water

 
Pedro Leite Ferreira and Dídia Isabel Cameira Covas    
The current paper aims at assessing the effect of the radial mesh on the description of the axial velocity in steady-state and transient conditions and at presenting the results of a new optimized equal-area mesh. For this purpose, a quasi-2D model is im... ver más
Revista: Water

 
Klemen Prah and Ashton M. Shortridge    
This paper investigated the surface length of roads from both horizontal and vertical perspectives using the theory of fractal dimension of surfaces and curves. Three progressive experiments were conducted. The first demonstrated the magnitude of the dif... ver más

 
Haotian You, Xu Tang, Qixu You, Yao Liu, Jianjun Chen and Feng Wang    
Light Detection and Ranging (LiDAR) points and high-resolution RGB image-derived points have been successfully used to extract tree structural parameters. However, the differences in extracting individual tree structural parameters among different tree s... ver más
Revista: Drones