Automatic Building Segmentation of Aerial Imagery Using Multi-Constraint Fully Convolutional Networks

Wu, Guoxing; Shao, Xiaowei; Guo, Zhiling; Chen, Qi; Yuan, Weiqi; Shi, Xiaodan; Xu, Yang; Shibasaki, Ryosuke

doi:10.3390/rs10030407

Cited by 180 publications

(120 citation statements)

References 45 publications

Supporting

Mentioning

103

Contrasting

Order By: Relevance

“…DeepResUnet adopted the U-Net as its basic structure and meanwhile took advantages of deep residual learning by replacing the plain neural units of the U-Net with the residual learning units to facilitate the training process of the network. Although some combined approaches of the U-Net with the deep residual network have been reported in recent studies [51,57,58], significant differences can be found between their network architectures and that of ours.…”

Section: About the Deepresunetmentioning

confidence: 60%

Semantic Segmentation of Urban Buildings from VHR Remote Sensing Imagery Using a Deep Convolutional Neural Network

Zhang

et al. 2019

Remote Sensing

169

View full text Add to dashboard Cite

Urban building segmentation is a prevalent research domain for very high resolution (VHR) remote sensing; however, various appearances and complicated background of VHR remote sensing imagery make accurate semantic segmentation of urban buildings a challenge in relevant applications. Following the basic architecture of U-Net, an end-to-end deep convolutional neural network (denoted as DeepResUnet) was proposed, which can effectively perform urban building segmentation at pixel scale from VHR imagery and generate accurate segmentation results. The method contains two sub-networks: One is a cascade down-sampling network for extracting feature maps of buildings from the VHR image, and the other is an up-sampling network for reconstructing those extracted feature maps back to the same size of the input VHR image. The deep residual learning approach was adopted to facilitate training in order to alleviate the degradation problem that often occurred in the model training process. The proposed DeepResUnet was tested with aerial images with a spatial resolution of 0.075 m and was compared in performance under the exact same conditions with six other state-of-the-art networks-FCN-8s, SegNet, DeconvNet, U-Net, ResUNet and DeepUNet. Results of extensive experiments indicated that the proposed DeepResUnet outperformed the other six existing networks in semantic segmentation of urban buildings in terms of visual and quantitative evaluation, especially in labeling irregular-shape and small-size buildings with higher accuracy and entirety. Compared with the U-Net, the F1 score, Kappa coefficient and overall accuracy of DeepResUnet were improved by 3.52%, 4.67% and 1.72%, respectively. Moreover, the proposed DeepResUnet required much fewer parameters than the U-Net, highlighting its significant improvement among U-Net applications. Nevertheless, the inference time of DeepResUnet is slightly longer than that of the U-Net, which is subject to further improvement.

show abstract

Section: About the Deepresunetmentioning

confidence: 60%

Semantic Segmentation of Urban Buildings from VHR Remote Sensing Imagery Using a Deep Convolutional Neural Network

Zhang

et al. 2019

Remote Sensing

169

View full text Add to dashboard Cite

show abstract

“…[25,26] trained FCNs to extract the buildings using the patch-wise method. In [27], Wu et al built a multi-constraint network to sharpen the boundaries of artificial object predictions. A trainable block, called the field-of-view (FoV), is proposed in [28] to boost the performance of the FCN.…”

mentioning

confidence: 99%

Web-Net: A Novel Nest Networks with Ultra-Hierarchical Sampling for Building Extraction from Aerial Imageries

et al. 2019

View full text Add to dashboard Cite

How to efficiently utilize vast amounts of easily accessed aerial imageries is a critical challenge for researchers with the proliferation of high-resolution remote sensing sensors and platforms. Recently, the rapid development of deep neural networks (DNN) has been a focus in remote sensing, and the networks have achieved remarkable progress in image classification and segmentation tasks. However, the current DNN models inevitably lose the local cues during the downsampling operation. Additionally, even with skip connections, the upsampling methods cannot properly recover the structural information, such as the edge intersections, parallelism, and symmetry. In this paper, we propose the Web-Net, which is a nested network architecture with hierarchical dense connections, to handle these issues. We design the Ultra-Hierarchical Sampling (UHS) block to absorb and fuse the inter-level feature maps to propagate the feature maps among different levels. The position-wise downsampling/upsampling methods in the UHS iteratively change the shape of the inputs while preserving the number of their parameters, so that the low-level local cues and high-level semantic cues are properly preserved. We verify the effectiveness of the proposed Web-Net in the Inria Aerial Dataset and WHU Dataset. The results of the proposed Web-Net achieve an overall accuracy of 96.97% and an IoU (Intersection over Union) of 80.10% on the Inria Aerial Dataset, which surpasses the state-of-the-art SegNet 1.8% and 9.96%, respectively; the results on the WHU Dataset also support the effectiveness of the proposed Web-Net. Additionally, benefitting from the nested network architecture and the UHS block, the extracted buildings on the prediction maps are obviously sharper and more accurately identified, and even the building areas that are covered by shadows can also be correctly extracted. The verified results indicate that the proposed Web-Net is both effective and efficient for building extraction from high-resolution remote sensing images.The building extraction task has drawn the attention of researchers over recent years. Before the common application of deep learning, there were massive machine learning models that tried to handle this task. In general, the pixel-wise labelling model consists of two sub-modules: the feature descriptor for extracting the semantic features from the original images and the pixel-wise classifier for determining the classes of the pixels. Some carefully designed feature descriptors were widely used in early approaches. Tuermer et al. [4] firstly used the histogram of gradient (HOG) feature descriptor in remote sensing for detecting vehicles. The Haar feature is applied in [5] for detecting buildings' outlines and determining the location of buildings' corners. Additionally, Yang et al. [6] applied the Scale-invariant feature transform [7] (SIFT) for classifying objects in remote sensing images. Unlike the artificially designed feature descriptors, the trainable models are the mainstream for the choices of classifiers....

show abstract

“…Facing this problem, we introduce Geoseg (https://github.com/ huster-wgm/geoseg), a computer vision package that is focus on implementing the state-of-the-art methods for automatic and accurate building segmentation and outline extraction. The Geoseg package implements more than 9 FCN-based models including FCNs [9], U-Net [10], SegNet [11], FPN [12], ResUNet [13], MC-FCN [14], and BR-Net [15]. For in-depth comparison, balanced and unbalanced evaluation metrics, such as precision, recall, overall accuracy, f1score, Jaccard index or intersection over union (IoU) [16] and kappa coefficient [17], are implemented.…”

Section: Introductionmentioning

confidence: 99%

GEOSEG: A Computer Vision Package for Automatic Building Segmentation and Outline Extraction

Guo

Shao

et al. 2019

IGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing Symposium

Self Cite

View full text Add to dashboard Cite

Recently, deep learning algorithms, especially fully convolutional network based methods, are becoming very popular in the field of remote sensing. However, these methods are implemented and evaluated through various datasets and deep learning frameworks. There has not been a package that covers these methods in a unifying manner. In this study, we introduce a computer vision package termed Geoseg that focus on building segmentation and outline extraction. Geoseg implements over nine state-of-the-art models as well as utility scripts needed to conduct model training, logging, evaluating and visualization. The implementation of Geoseg emphasizes unification, simplicity, and flexibility. The performance and computational efficiency of all implemented methods are evaluated by comparison experiment through a unified, high-quality aerial image dataset.

show abstract

Automatic Building Segmentation of Aerial Imagery Using Multi-Constraint Fully Convolutional Networks

Cited by 180 publications

References 45 publications

Semantic Segmentation of Urban Buildings from VHR Remote Sensing Imagery Using a Deep Convolutional Neural Network

Semantic Segmentation of Urban Buildings from VHR Remote Sensing Imagery Using a Deep Convolutional Neural Network

Web-Net: A Novel Nest Networks with Ultra-Hierarchical Sampling for Building Extraction from Aerial Imageries

GEOSEG: A Computer Vision Package for Automatic Building Segmentation and Outline Extraction

Contact Info

Product

Resources

About