Extracting ships from complex backgrounds is the bottleneck of ship detection in high-resolution optical satellite images. In this letter, we propose a nearly closed-form ship rotated bounding box space used for ship detection and design a method to generate a small number of highly potential candidates based on this space. We first analyze the possibility of accurately covering all ships by labeling rotated bounding boxes. Moreover, to reduce search space, we construct a nearly closed-form ship rotated bounding box space.

This dataset is the supporting simulated data for the paper titled "Hidden Border Tunnels: Research on Excavation Monitoring and Excavation Path Prediction." These data are generated through physical simulations and are used to validate the effectiveness of the algorithms proposed in the paper. The dataset includes the coordinates of simulated vibration events, as well as the prediction results of excavation events by various machine learning models, such as RNN, LSTM, GRU, Transformer, CNN_Transformer, and LSTMTransformer.

This dataset includes three types: DEM, land cover data, and high-definition remote sensing images.

Among them, the DEM data has an accuracy of 30 meters and is used to calculate the slope value. The land cover data has an accuracy of 8.98 meters and is located in a field area in southern Taiwan, China, and is used to extract coarse-grained land object types. The high-definition remote sensing data is taken by drones with an accuracy of 0.03 meters and is used for unsupervised classification of land objects.

MapData is a globally diverse dataset spanning 233 geographic sampling points. It offers original high-resolution images ranging from 7,000×5,000 to 20,000×15,000 pixels. After rigorous cleaning, the dataset provides 121,781 aligned electronic map–visible image pairs (each standardized to 512×512 pixels) with hybrid manual-automated ground truth—addressing the scarcity of scalable multimodal benchmarks.