Automatic extraction of buildings from remote sensing imagery plays a significant role in many applications, such as urban planning and monitoring changes to land cover. Various building segmentation methods have been proposed for visible remote sensing images, especially state-of-the-art methods based on convolutional neural networks (CNNs). However, high-accuracy building segmentation from high-resolution remote sensing imagery is still a challenging task due to the potentially complex texture of buildings in general and image background. Repeated pooling and striding operations used in CNNs reduce feature resolution causing a loss of detailed information. To address this issue, we propose a light-weight deep learning model integrating spatial pyramid pooling with an encoder-decoder structure. The proposed model takes advantage of a spatial pyramid pooling module to capture and aggregate multi-scale contextual information and of the ability of encoder-decoder networks to restore losses of information. The proposed model is evaluated on two publicly available datasets; the Massachusetts roads and buildings dataset and the INRIA Aerial Image Labeling Dataset. The experimental results on these datasets show qualitative and quantitative improvement against established image segmentation models, including SegNet, FCN, U-Net, Tiramisu, and FRRN. For instance, compared to the standard U-Net, the overall accuracy gain is 1.0% (0.913 vs. 0.904) and 3.6% (0.909 vs. 0.877) with a maximal increase of 3.6% in model-training time on these two datasets. These results demonstrate that the proposed model has the potential to deliver automatic building segmentation from high-resolution remote sensing images at an accuracy that makes it a useful tool for practical application scenarios.INDEX TERMS Deep learning, high-resolution remote sensing imagery, building extraction, fully convolutional networks, encoder-decoder.

Automatic building extraction based on high-resolution aerial images has important applications in urban planning and environmental management. In recent years advances and performance improvements have been achieved in building extraction through the use of deep learning methods. However, the design of existing models focuses attention to improve accuracy through an overflowing number of parameters and complex structure design, resulting in large computational costs during the learning phase and low inference speed. To address these issues, we propose a new, efficient end-to-end model, called ARC-Net. The model includes residual blocks with asymmetric convolution (RBAC) to reduce the computational cost and to shrink the model size. In addition, dilated convolutions and multi-scale pyramid pooling modules are utilized to enlarge the receptive field and to enhance accuracy. We verify the performance and efficiency of the proposed ARC-Net on the INRIA Aerial Image Labeling dataset and WHU building dataset. Compared to available deep learning models, the proposed ARC-Net demonstrates better segmentation performance with less computational costs. This indicates that the proposed ARC-Net is both effective and efficient in automatic building extraction from high-resolution aerial images.

No abstract

With the booming development of evacuation simulation software, developing an extensive database in indoor scenarios for evacuation models is imperative. In this paper, we conduct a qualitative and quantitative analysis of the collected videotapes and aim to provide a complete and unitary database of pedestrians’ earthquake emergency response behaviors in indoor scenarios, including human-environment interactions. Using the qualitative analysis method, we extract keyword groups and keywords that code the response modes of pedestrians and construct a general decision flowchart using chronological organization. Using the quantitative analysis method, we analyze data on the delay time, evacuation speed, evacuation route and emergency exit choices. Furthermore, we study the effect of classroom layout on emergency evacuation. The database for indoor scenarios provides reliable input parameters and allows the construction of real and effective constraints for use in software and mathematical models. The database can also be used to validate the accuracy of evacuation models.

Thin cirrus clouds are dominated by non-spherical ice crystals with an effective emissivity of less than 0.5. Until now, the influences of clouds were not commonly considered in the development of algorithms for retrieving land-surface temperature (LST). However, numerical simulations showed that the influence of thin cirrus clouds could lead to a maximum LST retrieval error of more than 14 K at night if the cirrus optical depth (COD) at 12 μm was equal to 0.7 (cirrus emissivity equivalent to 0.5). To obtain an accurate estimate of the LST under thin cirrus using satellite infrared data, a nonlinear three-channel LST retrieval algorithm was proposed based on a widely used two-channel algorithm for clear-sky conditions. The variations in the cloud top height, COD, and effective radius of cirrus clouds were considered in this three-channel LST retrieval algorithm. Using Moderate Resolution Imaging Spectroradiometer (MODIS) channels 20, 31, and 32 (centred at 3.8, 11.0, and 12.0 μm, respectively) and the corresponding land surface emissivities (LSEs), the simulated data showed that this algorithm could obtain LSTs with root mean square errors (RMSEs) of less than 2.8 K when the COD at 12 μm is less than 0.7 and the viewing zenith angle (VZA) is less than 60°. In addition, a sensitivity analysis of the proposed algorithm showed that the total LST errors, including errors from the uncertainties in input parameters and algorithm error, were nearly the same as the algorithm error itself. Some lake surface water temperatures measured in Lake Superior and Lake Erie were used to test the performance of the proposed LST retrieval algorithm. The results showed that the proposed nonlinear three-channel algorithm could be used for estimating LST under thin cirrus with an RMSE of less than 2.8 K.