With the advancement of mobile mapping technology, multi-view imagery has become an important source for road observation. However, traditional matching methods struggle to overcome the challenges posed by distortions and variations in viewing angles between images. To enhance the precision and reliability of image matching, this study investigates a deep feature matching technique based on deep learning. By utilizing convolutional neural networks (CNN) to extract deep features, accurate matching and 3D spatial positioning of multi-view images can be achieved. The study employs Deep Feature Matching (DFM) technology, which is based on the pre-trained VGG19 model. Through a two-stage matching strategy and the RANSAC (Random Sample Consensus) algorithm, erroneous matching points are filtered out to ensure the reliability of the matching results. The experimental data consists of multi-view images, with traffic signs serving as the target objects for image matching and positioning. The research results reveal that, compared with the traditional SIFT method, DFM demonstrates a higher success rate and improved positioning accuracy in various image scenarios, including scale differences, shape distortions, and occlusion conditions. Notably, DFM achieves significantly more matching points than SIFT in distortion and occlusion scenarios. Furthermore, the analysis of traffic sign positioning indicates that the success rate of traffic sign positioning reaches 70%, with an average error of less than 0.5 m for successfully located points. This finding highlights the practical application potential of DFM in 3D positioning using multi-view images in complex scenarios and confirms that it achieves higher success rates and accuracy.

The western plains of Taiwan have long relied on groundwater, and combined with uneven rainfall and droughts caused by climate change, land subsidence is easily triggered. Severe droughts in 2021 and 2023 further intensified subsidence during those years. This study applies the Persistent Scatterer Interferometric Synthetic Aperture Radar (PSInSAR) technique to analyze Sentinel-1 images from 2019 to 2023, integrated with rainfall data, to examine the impact of drought on subsidence rates. Results indicate that precipitation decreased sharply during drought periods and showed strong correlation with accelerated subsidence, with seasonal subsidence rates reaching up to 7.8 cm/year in severely water-stressed areas. To validate the results, GNSS stations were incorporated using multi-reference kriging interpolation, with independent stations used for comparison, yielding a correlation coefficient of 0.98. Furthermore, comparisons with significant subsidence areas published by the Water Resources Agency revealed differences of about ±10% between 2020 and 2023. These findings demonstrate that PSInSAR provides high accuracy and stability in monitoring detailed surface deformation and can serve as a valuable scientific basis for future water resource management, groundwater regulation, and land subsidence prevention.

Taiwan's unique topography and environmental conditions make it highly susceptible to natural disasters, which often result in significant environmental damage and pose threats to human lives and property. Given the urgent nature of disaster response, it is imperative to obtain timely and accurate information regarding affected areas. Furthermore, the risk of secondary disasters must not be underestimated, underscoring the importance of comprehensive spatial data collection in disaster-stricken regions. The completeness and accuracy of such data directly influence the efficiency and effectiveness of emergency response and recovery operations. This study presents a case in which a combination of remote sensing technologies—including Unmanned Aerial Vehicles (UAVs) for aerial imagery and 3D reconstruction, Light Detection and Ranging (LiDAR) for point cloud modeling, and total stations for geodetic surveying—was employed to rapidly acquire spatial information in disaster areas. The integrated approach provided critical support for emergency repair and post-disaster reconstruction, demonstrating the value of adopting diverse technologies to meet varying operational requirements.

Road construction often results in utility damage due to inaccurate mapping data, primarily because contractors submit design drawings as-built drawings. This leads to discrepancies in pipeline spatial location and burial depth, as the submitted data does not accurately reflect on-site conditions. To address this issue, the Taoyuan Aerotropolis Engineering Office has implemented a 3D Geographic Information System (3D GIS) that integrates BIM design models, surveyed pipeline points, and data in Geography Markup Language (GML) format to enhance the accuracy of utility mapping. To ensure data quality, the as-built drawings must pass a system-based rationality verification process. In addition, supervising engineers manually inspect the pipeline data for continuity, completeness, and accuracy. The validated data is then integrated into a 3D public utility database to improve overall mapping quality, meet future utility maintenance needs within the Aerotropolis project area, and reduce the risk of utility damage and associated threats to life and property during excavation.

The importance of urban water resource management is increasing, requiring advanced tools to support operations and decision-making. This study integrates artificial intelligence, hydraulic modeling, and geographic information systems to enhance analysis and overcome the limitations of traditional approaches in assessing affected users. Focusing on Taipei’s water distribution network, the framework applies three key practices. First, construction impact analysis identifies service areas and households affected by pipeline replacement, enabling accurate planning and communication. Second, network planning and design use smart isolation techniques by controlling boundary valves to determine the optimal inflow point and improve water distribution efficiency. Third, multi-period source tracing applies dynamic hydraulic modeling to quantify water source proportions at each node and track major flow paths across time. This integrated methodology provides a powerful means to dynamically evaluate user impacts and precisely map results to individual pipelines. For the Taipei Water Department, it offers tailored management strategies that improve maintenance efficiency, strengthen decision support for pipeline renewal, and advance sustainable, data-driven water management.

In order to align with the central government's policy on land Information System while to meet the needs of the county government’s development and surveying in both official and private department, Taitung County Government promote the project of "High-Precision Digital Terrain Map Data Refinement and 3D Mesh in Ludao Township". This project deploy the surveying aircraft equipped with LiDAR system and nadir/oblique camera system. The result of this project includes 6 cm True Orthophotos, 10 sites of LOD3 mesh, 1/1000 mapping using the 6 cm True Orthophotos, which is the first attempt in nation, and the accuracy of these results exceeds standards, achieving the goal of "image as map". This project aims to transform Green Island into an experimental area of digital twin and 3D GIS. The high-precision, high-quality data produced can be applied to cadastral surveying, land planning, civil construction, tourist routes, and disaster prevention. It is sure to become the best tool for Taitung County Government in planning, evaluation, demonstration, and execution.