Spatial population distribution data is the discretization of demographic data into spatial grids, which has vital reference significance for disaster emergency response, disaster assessment, emergency rescue resource allocation, and post-disaster reconstruction. The random forest (RF) model, as a prominent method for modeling the spatial distribution of population, has been studied by many scholars, both domestically and abroad. Specifically, research has focused on aspects such as multi-source data fusion, feature selection, and data accuracy evaluation within the modeling process. However, discussions about parameter optimization methods during the modeling process and the impact of different optimization methods on modeling accuracy are relatively limited. In light of the above circumstances, this paper employs the RF model to conduct research on population spatialization with multi-source spatial information data. The study primarily explores the differences in model parameter optimization achieved through random search algorithms, grid search algorithms, genetic algorithms, simulated annealing algorithms, Bayesian optimization based on Gaussian process algorithms, and Bayesian optimization based on gradient boosting regression tree algorithms. Additionally, the study investigates the influence of different optimization algorithms on the accuracy of population spatialization modeling. Subsequently, the model with the highest accuracy is selected as the prediction model for population spatialization. Based on this model, a spatial population distribution dataset of Sichuan Province at a 1 km resolution is generated. Finally, the population dataset created in this paper is compared and validated with open datasets such as GPW, LandScan, and WorldPop. Experimental results indicate that the spatial population distribution dataset produced by the Bayesian optimization-based random forest model proposed in this paper exhibits a higher fitting accuracy with real data. The Coefficient of Determination (R²) is 0.6628, the Mean Absolute Error (MAE) is 12,459, and the Root Mean Squared Error (RMSE) is 25,037. Compared to publicly available international datasets, the dataset generated in this paper more accurately represents the spatial distribution of the population.

Unmanned aerial vehicle (UAV) remote sensing plays a crucial role in the rapid assessment of post-earthquake damaged buildings. Existing methods for automatic identification of damaged buildings based on UAV remote sensing imagery primarily focus on assessment accuracy. However, during earthquake emergency response, both speed and accuracy of assessment are equally important. Therefore, we propose a high-precision and real-time damaged building instance segmentation model, DBA-RTMDet, based on the RTMDet framework. Three key modules are designed to enhance the network's feature extraction capability for damaged buildings. First, we develop a Depthwise Separable Atrous Spatial Pyramid Pooling Module to improve the model's ability to detect damaged buildings of varying sizes. Second, an Adjacent Level Feature Enhancement and Fusion Module is introduced to enhance the network's ability to localize the irregular boundaries of collapsed buildings. Finally, to mitigate the interference caused by intact buildings that exhibit features like partially damaged roofs, we incorporated a Cross Atrous Attention Module. We conduct experiments on post-disaster UAV imagery collected during the 2021 Yangbi earthquake with surface wave magnitude scale (Ms) of 6.4 in Yunnan, China. The results demonstrate that our model achieves a segmentation mean Average Precision (mAP^seg) of 80.4% and a bounding box mean Average Precision (mAP^box) of 81.2%, significantly outperforming methods such as YOLACT, FCOS, PolarMask, TensorMask, YOLOv5, YOLOX, YOLOv6, and YOLOv8. We observed that, compared to the widely used model YOLOv5, DBA-RTMDet not only improves mAP^seg and mAP^box by 10.3% and 9.7%, respectively, but also the Frames Per Second (FPS) is increased significantly by 26. Our approach can quickly and accurately identify damaged buildings, with its high application value verified through the generalization validation experiment. The code and datasets will be available at https://github.com/Kangsan-Y/UAV-based-Damaged-Buildings-Instance-Segmentation-model-DBA-RTMDet.

Automatic extraction of building contours from high-resolution images is of great significance in the fields of urban planning, demographics, and disaster assessment. Network models based on convolutional neural network (CNN) and transformer technology have been widely used for semantic segmentation of buildings from high resolution remote sensing images (HRSI). However, the fixed geometric structure and the local receptive field of the convolutional kernel are not good at global feature extraction, and the transformer technique with self-attention mechanism introduces computational redundancies and extracts local feature details poorly in the process of modeling the global contextual information. In this paper, a dual-branch fused reconstructive transformer network, DFRTNet, is proposed for efficient and accurate building extraction. In the encoder, the traditional transformer is reconfigured by designing the local and global feature extraction module (LGFE); the branch of global feature extraction (GFE) performs dynamic range attention (DRA) based on the idea of top-k attention for extracting global features; furthermore, the branch of local feature extraction (LFE) is used to obtain fine-grained features. The multilayer perceptron (MLP) is employed to efficiently fuse the local and global features. In the decoder, a simple channel attention module (CAM) is used in the up-sampling part to enhance channel dimension features. Our network achieved the best segmentation accuracy on both the WHU and Massachusetts building datasets when compared to other mainstream and state-of-the-art methods.

As unmanned aerial vehicle (UAV) remote sensing is applied in small area aerial photogrammetry surveying, disaster monitoring and emergency command, 3D urban construction and other fields, the image processing of UAV has become a hot topic in current research. The precise matching of UAV image is a key problem, which affects the subsequent processing precision directly, such as 3D reconstruction and automatic aerial triangulation, etc. At present, SIFT (Scale Invariant Feature Transform) algorithm proposed by DAVID G. LOWE as the main method is, is widely used in image matching, since its strong stability to image rotation, shift, scaling, and the change of illumination conditions. It has been successfully applied in target recognition, SFM (Structure from Motion), and many other fields. SIFT algorithm needs the colour images to be converted into grayscale images, detects extremum points under different scales and uses neighbourhood pixels to generate descriptor. As we all know that UAV images with rich colour information, the SIFT algorithm improved through combining with the image colour information in this paper, the experiments are conducted from matching efficiency and accuracy compared with the original SIFT algorithm. The results show that the method which proposed in this paper decreases on the efficiency, but is improved on the precision and provides a basis choice for matching method.

The image collection system based on the unmanned aerial vehicle plays an important role in the postearthquake response and disaster investigation. In the postearthquake response period, for hundreds of image stitching or 3D model reconstruction, the traditional UAV image processing methods may take one or several hours, which need to be improved on the efficiency. To solve this problem, the UAV image rapid georeference method for postearthquake is proposed in this paper. Firstly, we discuss the rapid georeference model of UAV images and then adopt the world file designed and developed by ESRI to organize the georeferenced image data. Next, the direct georeference method based on the position and attitude data collected by the autopilot system is employed to compute the upper-left corner coordinates of the georeferenced images. For the differences of image rotation manners between the rapid georeference model and the world file, the rapid georeference error compensation model from the image rotation is considered in this paper. Finally, feature extraction and feature matching for UAV images and referenced image are used to improve the accuracy of the position parameters in the world file, which will reduce the systematic error of the georeferenced images. We use the UAV images collected from Danling County and Beichuan County, Sichuan Province, to implement the rapid georeference experiments employing different types of UAV. All the images are georeferenced within three minutes. The results show that the algorithm proposed in this paper satisfies the time and accuracy requirements of postearthquake response, which has an important application value.

Unmanned aerial vehicle (UAV) remote sensing is critical in assessing post-earthquake building damage. However, intelligent disaster assessment via remote sensing faces formidable challenges from complex backgrounds, substantial scale variations in targets, and diverse spatial disaster dynamics. To address these issues, we propose UEBNet, a high-precision post-earthquake building instance segmentation model that systematically enhances damage recognition by integrating three key modules. Firstly, the Depthwise Separable Convolutional Block Attention Module suppresses background noise that visually resembles damaged structures. This is achieved by expanding the receptive field using multi-scale pooling and dilated convolutions. Secondly, the Multi-feature Fusion Module generates scale-robust feature representations for damaged buildings with significant size differences by processing feature streams from different receptive fields in parallel. Finally, the Adaptive Multi-Scale Interaction Module accurately reconstructs the irregular contours of damaged buildings through an advanced feature alignment mechanism. Extensive experiments were conducted using UAV imagery collected after the Ms 6.8 earthquake in Tingri County, Tibet Autonomous Region, China, on 7 January 2025, and the Ms 6.2 earthquake in Jishishan County, Gansu Province, China, on 18 December 2023. Results indicate that UEBNet enhances segmentation mean Average Precision (${mAP}^{seg}$) and bounding box mean Average Precision (${mAP}^{box}$) by 3.09% and 2.20%, respectively, with equivalent improvements of 2.65% in F1-score and 1.54% in overall accuracy, outperforming state-of-the-art instance segmentation models. These results demonstrate the effectiveness and reliability of UEBNet in accurately segmenting earthquake-damaged buildings in complex post-disaster scenarios, offering valuable support for emergency response and disaster relief.