Fully convolutional networks for building and road extraction: Preliminary results

Zhong, Zilong; Li, Jonathan; Cui, Weihong; Han, Jiang

doi:10.1109/igarss.2016.7729406

Cited by 145 publications

(82 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the remote sensing area, according to the mentioned properties of high-resolution remote sensing images, some carefully designed models have been proposed and optimized for building extraction tasks that are based on these above semantic segmentation approaches. In early research, [23,24] used naive FCN architectures with deconvolutional layers to extract buildings or roads, and these works demonstrated the effectiveness and efficiency of the FCN architecture. [25,26] trained FCNs to extract the buildings using the patch-wise method.…”

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

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

show abstract

“…However, in machine learning, at least three major factors that influence accuracy of classification should be considered. Especially, selection of training samples, feature selection and settings of tuning parameters [7]. Although, the three aforementioned factors have been investigated in the past, only few investigations in relation to settings of tuning parameters have been witnessed.…”

Section: Introductionmentioning

confidence: 99%

Using of machines learning in extraction of urban roads from DEM of LIDAR data: Case study at Baghdad expressways, Iraq

Jasim¹

2019

PEN

View full text Add to dashboard Cite

Road extraction from remotely perceived information may be a difficult issue and has been approached in many various ways in which by photogrammetrists and digital image processors. This study was given extraction of roads from DEM of LiDAR with IKONOS image using machines learning (ML). Two set of data were used in this study. IKONOS Image and Digital Elevation Model (DEM) data were combined to produce thematic mapping. The spatial resolution of data is 1 m and was acquired on 2010. The result f thematic map based on theses images and the methods was used three models of machines learning. The problem of this study, when was used the LIDAR data to extract the road is very difficult because the LiDAR data is too noisy and employed it also so hard. Moreover, this article will describe an effective and compare between several machines learning algorithms (RF, BYO and MLP) for detection the roads from LIDAR data. The statistical indictors such as an overall accuracy (OA), kappa analysis statistic (K), (MAE) which is Mean Absolute Error and finally the (RMSE) which is Root Mean Squared Error. All these will be use to get the accuracy of classification assessment and the best model to produce the thematic map.

show abstract

“…In the face of these challenges, recent studies have tried to apply supervised deep learning (DL) models to extract robust and discriminant features in the context of remotely sensed image classification [3,4]. In 2016, convolutional neural network (CNN) was used to extract spatial features that integrated with spectral features learned from a embedding method [5].…”

Section: Introductionmentioning

confidence: 99%

Deep residual networks for hyperspectral image classification

Zhong

Han

et al. 2017

2017 IEEE International Geoscience and Remote Sensing Symposium (IGARSS)

Self Cite

View full text Add to dashboard Cite

Deep neural networks can learn deep feature representation for hyperspectral image (HSI) interpretation and achieve high classification accuracy in different datasets. However, counterintuitively, the classification performance of deep learning models degrades as their depth increases. Therefore, we add identity mappings to convolutional neural networks for every two convolutional layers to build deep residual networks (ResNets). To study the influence of deep learning model size on HSI classification accuracy, this paper applied ResNets and CNNs with different depth and width using two challenging datasets. Moreover, we tested the effectiveness of batch normalization as a regularization method with different model settings. The experimental results demonstrate that ResNets mitigate the declining-accuracy effect and achieved promising classification performance with 10% and 5% training sample percentages for the University of Pavia and Indian Pines datasets, respectively. In addition, t-Distributed Stochastic Neighbor Embedding (t-SNE) provides a direct view of the extracted features through dimensionality reduction.

show abstract

Fully convolutional networks for building and road extraction: Preliminary results

Cited by 145 publications

References 9 publications

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

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

Using of machines learning in extraction of urban roads from DEM of LIDAR data: Case study at Baghdad expressways, Iraq

Deep residual networks for hyperspectral image classification

Contact Info

Product

Resources

About