Feature-based matching methods have been widely used in remote sensing image matching given their capability to achieve excellent performance despite image geometric and radiometric distortions. However, most of the feature-based methods are unreliable for complex background variations, because the gradient or other image grayscale information used to construct the feature descriptor is sensitive to image background variations. Recently, deep learning-based methods have been proven suitable for high-level feature representation and comparison in image matching. Inspired by the progresses made in deep learning, a new technical framework for remote sensing image matching based on the Siamese convolutional neural network is presented in this paper. First, a Siamese-type network architecture is designed to simultaneously learn the features and the corresponding similarity metric from labeled training examples of matching and non-matching true-color patch pairs. In the proposed network, two streams of convolutional and pooling layers sharing identical weights are arranged without the manually designed features. The number of convolutional layers is determined based on the factors that affect image matching. The sigmoid function is employed to compute the matching and non-matching probabilities in the output layer. Second, a gridding sub-pixel Harris algorithm is used to obtain the accurate localization of candidate matches. Third, a Gaussian pyramid coupling quadtree is adopted to gradually narrow down the searching space of the candidate matches, and multiscale patches are compared synchronously. Subsequently, a similarity measure based on the output of the sigmoid is adopted to find the initial matches. Finally, the random sample consensus algorithm and the whole-to-local quadratic polynomial constraints are used to remove false matches. In the experiments, different types of satellite datasets, such as ZY3, GF1, IKONOS, and Google Earth images, with complex background variations are used to evaluate the performance of the proposed method. The experimental results demonstrate that the proposed method, which can significantly improve the matching performance of multi-temporal remote sensing images with complex background variations, is better than the state-of-the-art matching methods. In our experiments, the proposed method obtained a large number of evenly distributed matches (at least 10 times more than other methods) and achieved a high accuracy (less than 1 pixel in terms of root mean square error).

Earthquake-triggered landslides (ETLs) are characterized by their extensive occurrence,having wide distributions. The conventional human-computer interaction extraction method is often time-consuming and labor-intensive, failing to meet the demands of disaster emergency response. There is a pressing need for a swift detection of ETLs. In this study, we introduce an earthquake triggered landslides extraction method (MAST) combining morphological analysis edge recognition with Swin-Transformer deep learning model which is specifically designed for landslide extraction. The MAST model adopts a hierarchical construction approach akin to convolution neural networks, aiding in tasks such as target detection and semantic segmentation. To enhance the accuracy of landslide edge extraction, we incorporate an edge recognition algorithm based on morphological analysis into the MAST model. This algorithm leverages morphological operations to extract the features of landslide boundaries. It effectively addresses issues such as discretization and irregularization of the extracted landslide boundaries, leading to more precise delineation of landslide boundaries.Drawing on UAV data collected from Wan Dong Village, De Tou Town, Sichuan Luding, China, during the 2022 Ms 6.8 Luding Earthquake, we conduct automated extraction of ETLs utilizing the MAST model. Experimental results demonstrate the superior performance of the MAST model compared to the traditional full convolution neural network model (FCN)and normal Swin-Transformer model (SWT). The MAST model exhibits enhanced most value in landslide extraction. Notably, it demonstrates a significant advantage in boundary extraction. Employing the Boundary IoU metric to evaluate the accuracy of ETLs extraction, the MAST model outperforms the SWT and FCN models at various distances.

In this study, an approach using ground control point-free unmanned aerial vehicle (UAV)-based photogrammetry is proposed to estimate the volume of stockpiles carried on barges in a dynamic environment. Compared with similar studies regarding UAVs, an indirect absolute orientation based on the geometry of the vessel is used to establish a custom-built framework that can provide a unified reference instead of prerequisite ground control points (GCPs). To ensure sufficient overlap and reduce manual intervention, the stereo images are extracted from a UAV video for aerial triangulation. The region of interest is defined to exclude the area of water in all UAV images using a simple linear iterative clustering algorithm, which segments the UAV images into superpixels and helps to improve the accuracy of image matching. Structure-from-motion is used to recover three-dimensional geometry from the overlapping images without assistance of exterior parameters obtained from the airborne global positioning system and inertial measurement unit. Then, the semi-global matching algorithm is used to generate stockpile-covered and stockpile-free surface models. These models are oriented into a custom-built framework established by the known distance, such as the length and width of the vessel, and they do not require GCPs for coordinate transformation. Lastly, the volume of a stockpile is estimated by multiplying the height difference between the stockpile-covered and stockpile-free surface models by the size of the grid that is defined using the resolution of these models. Results show that a relatively small deviation of approximately ±2% between the volume estimated by UAV photogrammetry and the volume calculated by traditional manual measurement was obtained. Therefore, the proposed approach can be considered the better solution for the volume measurement of stockpiles carried on barges in a dynamic environment because UAV-based photogrammetry not only attains superior density and spatial object accuracy but also remarkably reduces data collection time.

Improving the matching reliability of multi-sensor imagery is one of the most challenging issues in recent years, particularly for synthetic aperture radar (SAR) and optical images. It is difficult to deal with the noise influence, geometric distortions, and nonlinear radiometric difference between SAR and optical images. In this paper, a method for SAR and optical images matching is proposed. First, interest points that are robust to speckle noise in SAR images are detected by improving the original phase-congruency-based detector. Second, feature descriptors are constructed for all interest points by combining a new Gaussian-Gamma-shaped bi-windows-based gradient operator and the histogram of oriented gradient pattern. Third, descriptor similarity and geometrical relationship are combined to constrain the matching processing. Finally, an approach based on global and local constraints is proposed to eliminate outliers. In the experiments, SAR images including COSMO-Skymed, RADARSAT-2, TerraSAR-X and HJ-1C images, and optical images including ZY-3 and Google Earth images are used to evaluate the performance of the proposed method. The experimental results demonstrate that the proposed method provides significant improvements in the number of correct matches and matching precision compared with the state-of-the-art SIFT-like methods. Near 1 pixel registration accuracy is obtained based on the matching results of the proposed method.

Tree heights are the principal variables for forest plantation inventory. The increasing availability of high-resolution three-dimensional (3D) point clouds derived from low-cost Unmanned Aerial Vehicle (UAV) and modern photogrammetry offers an opportunity to generate a Canopy Height Model (CHM) in the mountainous areas. In this paper, we assessed the capabilities of tree height estimation using UAV-based Structure-from-Motion (SfM) photogrammetry and Semi-Global Matching (SGM). The former is utilized to generate 3D geometry, while the latter is used to generate dense point clouds from UAV imagery. The two algorithms were coupled with a Radial Basis Function (RBF) neural network to acquire CHMs in mountainous areas. This study focused on the performance of Digital Terrain Model (DTM) interpolation over complex terrains. With the UAV-based image acquisition and image-derived point clouds, we constructed a 5 cm-resolution Digital Surface Model (DSM), which was assessed against 14 independent checkpoints measured by a Real-Time Kinematic Global Positioning System RTK GPS. Results showed that the Root Mean Square Errors (RMSEs) of horizontal and vertical accuracies are approximately 5 cm and 10 cm, respectively. Bare-earth Index (BEI) and Shadow Index (SI) were used to separate ground points from the image-derived point clouds. The RBF neural network coupled with the Difference of Gaussian (DoG) was exploited to provide a favorable generalization for the DTM from 3D ground points with noisy data. CHMs were generated using the height value in each pixel of the DSM and by subtracting the corresponding DTM value. Individual tree heights were estimated using local maxima algorithm under a contour-surround constraint. Two forest plantations in mountainous areas were selected to evaluate the accuracy of estimating tree heights, rather than field measurements. Results indicated that the proposed method can construct a highly accurate DTM and effectively remove nontreetop maxima. Furthermore, the proposed method has been confirmed to be acceptable for tree height estimation in mountainous areas given the strong linear correlation of the measured and estimated tree heights and the acceptable t-test values. Overall, the low-cost UAV-based photogrammetry and RBF neural network can yield a highly accurate DTM over mountainous terrain, thereby making them particularly suitable for rapid and cost-effective estimation of tree heights of forest plantation in mountainous areas.

Landcover classification is an important application in remote sensing, but it is always a challenge to distinguish different features with similar characteristics or large-scale differences. Some deep learning networks, such as UperNet, PSPNet, and DANet, use pyramid pooling and attention mechanisms to improve their abilities in multi-scale features extraction. However, due to the neglect of low-level features contained in the underlying network and the information differences between feature maps, it is difficult to identify small-scale objects. Thus, we propose a novel image segmentation network, named HFENet, for mining multi-level semantic information. Like the UperNet, HFENet adopts a top-down horizontal connection architecture while includes two improved modules, the HFE and the MFF. According to the characteristics of different levels of semantic information, HFE module reconstructs the feature extraction part by introducing an attention mechanism and pyramid pooling module to fully mine semantic information. With the help of a channel attention mechanism, MFF module up-samples and re-weights the feature maps to fuse them and enhance the expression ability of multi-scale features. Ablation studies and comparative experiments between HFENet and seven state-of-the-art models (U-Net, DeepLabv3+, PSPNet, FCN, UperNet, DANet and SegNet) are conducted with a self-labeled GF-2 remote sensing image dataset (MZData) and two open datasets landcover.ai and WHU building dataset. The results show that HFENet on three datasets with six evaluation metrics (mIoU, FWIoU, PA, mP, mRecall and mF1) are better than the other models and the mIoU is improved 7.41–10.60% on MZData, 1.17–11.57% on WHU building dataset and 0.93–4.31% on landcover.ai. HFENet can perform better in the task of refining the semantic segmentation of remote sensing images.