Building damage accounts for a high percentage of post-natural disaster assessment. Extracting buildings from optical remote sensing images is of great significance for natural disaster reduction and assessment. Traditional methods mainly are semi-automatic methods which require human-computer interaction or rely on purely human interpretation. In this paper, inspired by the recently developed deep learning techniques, we propose an improved Mask Region Convolutional Neural Network (Mask R-CNN) method that can detect the rotated bounding boxes of buildings and segment them from very complex backgrounds, simultaneously. The proposed method has two major improvements, making it very suitable to perform building extraction task. Firstly, instead of predicting horizontal rectangle bounding boxes of objects like many other detectors do, we intend to obtain the minimum enclosing rectangles of buildings by adding a new term: the principal directions of the rectangles θ. Secondly, a new layer by integrating advantages of both atrous convolution and inception block is designed and inserted into the segmentation branch of the Mask R-CNN to make the branch to learn more representative features. We test the proposed method on a newly collected large Google Earth remote sensing dataset with diverse buildings and very complex backgrounds. Experiments demonstrate that it can obtain promising results.

Disaster damage assessment is an important basis for the objective assessment of the social impacts of disasters and for the planning of recovery and reconstruction. It is also an important research field with regard to disaster mitigation and risk management. Quantitative assessment of physical damage refers to the determination of the physical damage state of the exposed elements in a disaster area, reflecting the aggregate quantities of damages. It plays a key role in the comprehensive damage assessment of major natural hazard-induced disasters. The National Disaster Reduction Center of China has established a technical work flow for the quantitative assessment of disaster physical damage using remote sensing data. This article presents a quantitative assessment index system and method that can be integrated with high-resolution remote sensing data, basic geographical data, and field survey data. Following the 2014 Ludian Earthquake in Yunnan Province, China, this work flow was used to assess the damage to buildings, roads, and agricultural and forest resources, and the assessment results were incorporated into the Disaster Damage Comprehensive Assessment Report of the 2014 Ludian Earthquake for the State Council of China. This article also outlines some possible improvements that can be addressed in future work.

Building structural types (BSTs) information is vital for seismic risk and vulnerability modeling. However, obtaining this kind of information is not a trivial task. The conventional method involves a labor-intensive and inefficient manual inspection process for each building. Nowadays, a few methods have explored to use remote sensing images and some building-related knowledge (BRK) to realize automated BSTs recognition. However, these methods have many limitations, such as insufficient mining of multimodal information and difficulty obtaining BRK, which hinders their promotion and practical use. To alleviate the shortcomings above, we propose a deep multimodal fusion model, which combines satellite optical remote sensing image, aerial Synthetic Aperture Radar (SAR) image, and BRK (roof type, color, and group pattern) obtained by domain experts to achieve accurate automatic reasoning of BSTs. Specifically, first, we use a pseudo-siamese network to extract the image feature. Second, a knowledge graph (KG) based on the BRK is constructed, and then we use a Graph Attention Network (GAT) to extract the semantic feature from the KG. Third, we propose a novel multi-stage gated fusion mechanism to fuse the image and semantic feature. Our method's best overall accuracy (OA) and kappa coefficient on the dataset collected in the study area are 90.35% and 0.83, which outperforms a series of existing methods. Through our model, high-precision BSTs information can be obtained for earthquake disaster prevention, reduction, and emergency decision-making.

Building detection from panchromatic (PAN) and Multi-spectral (MS) images is an essential task for many practical applications. In this paper, a dual-stream asymmetric fusion network is proposed, named DAFNet. DAFNet can achieve effective information fusion at the feature level. It obtains better building detection performance from three perspectives: 1) A two-stream network structure is designed to guarantee the ability to extract information from PAN and MS images; 2) An asymmetric feature fusion (AFF) module is proposed to fuse features efficiently and concisely; 3) Two consistency regularization losses, i.e., PAN information preservation (PiP) loss and cross-modal semantic consistency (CSC) loss are applied to further explore the consistency between features for better fusion. The experiments are conducted on a challenging building detection dataset collected from GaoFen-2 satellite images. Comprehensive evaluations on 12 popular detection methods demonstrate the superiority of our DAFNet compared with the existing state-of-the-art fusion methods. We reveal that feature-level fusion is more suitable for building detection from PAN-MS images. Code is released at https://github.com/floatingstarZ/DAFNet

In this study, modified perpendicular drought index (MPDI) models based on the red-near infrared spectral space are established for the first time through the analysis of the spectral characteristics of GF-1 wide field view (WFV) data, with a high spatial resolution of 16 m and the highest frequency as high as once every 4 days. GF-1 data was from the Chinese-made, new-generation high-resolution GF-1 remote sensing satellites. Soil-type spatial data are introduced for simulating soil lines in different soil types for reducing errors of using same soil line. Multiple vegetation indices are employed to analyze the response to the MPDI models. Relative soil moisture content (RSMC) and precipitation data acquired at selected stations are used to optimize the drought models, and the best one is the Two-band enhanced vegetation index (EVI2)-based MPDI model. The crop area that was statistically significantly affected by drought from a local governmental department, and used for validation. High correlations and small differences in drought-affected crop area was detected between the field observation data from the local governmental department and the EVI2-based MPDI results. The percentage of bias is between −21.8% and 14.7% in five sub-areas, with an accuracy above 95% when evaluating the performance via the data for the whole study region. Generally the proposed EVI2-based MPDI for GF-1 WFV data has great potential for reliably monitoring crop drought at a relatively high frequency and spatial scale. Currently there is almost no drought model based on GF-1 data, a full exploitation of the advantages of GF-1 satellite data and further improvement of the capacity to observe ground surface objects can provide high temporal and spatial resolution data source for refined monitoring of crop droughts.