CORE

遙感探測、智慧國土、空間資訊跨域應用

--- Research fields​ ---

​Remote Sensing; Intelligence Data Processing; Smart Homeland; Spatial Information Analysis; Interdisciplinary Smart Disaster Reduction; Spatiotemporal BIM Applications.

​

--- The objective ---

Applying spatial information derived from image processing and 3D mapping techniques to construction applications with a focus on promoting the level of visualization and automation.

Moderator

Tzu-Yi Chuang Ph.D.

Feb. 2021 -

Assistant Professor

Department of Civil Engineering

National Yang Ming Chiao Tung University

​

Aug. 2017 - Jan. 2021

Assistant Professor

Department of Civil and Construction Engineering

National Taiwan University of Science and Technology

Major Field

Spatial information for Engineering Applications

Remote Sensing

Photogrammetric Computer Vision

Learning-guided Methods

Interdisciplinary Spatial Information Analysis

BIM Applications

Courses

Construction Surveying

Applications of Geospatial Technology in Engineering

Intelligent Image Processing and 3D Mapping

Ongoing Projects

Stay in the Know

Inclination monitoring and management system development for offshore wind turbine towers

​

離岸風機塔柱傾斜檢測與管理系統開發

台灣具備眾多優良的潛力風場，經濟部風力發電計畫後續將投入更多深海場域與大型風機建設，如何精確地完成塔柱設置以及長期監測與維運管理顯得格外重要。然而，目前國內施工能量不足尚無法因應未來大型風機建置與維護需求，若長期租用國外施工船隊除了成本高昂外，亦會限制國內產業鏈之發展。風機塔柱的垂直度為施工安全與維運階段的重要指標，如何在海上不穩定的環境進行精確、有效且安全的監檢測為本計畫之重點。近年陸域的移動式測繪技術逐漸成熟，唯海域測繪相關技術門檻較高，使得商業化軟硬體設備高昂而無法落實長期離岸風機塔柱的檢測管理。本計畫擬開發一套成本合理且可靠的離岸風機塔柱傾斜檢測與管理系統，透過引入先進空間資訊學理以及適切的現代感測設備，規劃風機塔柱垂直度之檢測機制，並以移動式感測獲得塔柱多視角全方向資訊來建構單一整合式檢測與管理架構，並開發測繪資料處理與塔柱資訊整合分析軟體，提出國內自主研發的離岸風機塔柱例行性檢測方案。透過本計畫之執行，預期將可整合學術研發能力以及產業實務需求，在合理成本的條件下提升應用效能，不僅有助益於國內離岸風電再生能源政策的推動與相關產業的發展，同時提升國內技術研發的自主性。

​

​The offshore wind market is developing rapidly, the ministry of economic affairs of Taiwan keeps developing wind farm development in Taiwan strait. More deepsea wind farms and larger wind turbines will be constructed before 2025 referring to renewable energy policy. Therefore, how to overcome critical challenges to realize the installation and maintenance of wind towers in safety, accuracy, and sustainable way is of importance. Yet, the domestic industry chain currently lacks operation supports or specific vessels for offshore operations, which leading to huge costs for outsourcing solutions and losing technical autonomy. Considering the verticality of wind turbines is a significant index while driving foundations into the seafloor and evaluating structure stability, this project focuses on how to realize the effective and safe measurement of tower verticality in varied and unstable offshore conditions. Mobile mapping techniques have become viable solutions for land-based engineering. By contrast, the technical threshold for operations on the sea is relatively high and resulting in costly commercial equipment and making periodic monitoring and management flounder. This project will develop an offshore-use mobile monitoring system for wind tower verticality inspection and management and emphasize its affordability and effectiveness. By leveraging spatial information technology and proper sensors, the scheme of the single system can be drawn up and used to collect multi-view information. Also, the system realizes the mapping data processing and integrated analysis of tower information. Notably, the proposed system is domestic independent research and development and can be deemed as an alternative solution to periodic inspection for both offshore and land-based wind towers.

Applying Visual SLAM-based Indoor LiDAR Point Cloud Recognition and Change Detection Techniques to BIM-based Facility Management

​

運用Visual SLAM室內光達點雲辨識和變異偵測技術協助BIM設施管理

Recently, many studies have attempted to apply Building Information Modeling (BIM) to the facility management (FM) during the operational phase of a building life cycle to improve automation and effectiveness. However, a mechanism is currently lack for updating BIM models at each phase of a building life cycle, leading to errors or inaccurate data between BIM models and their actual conditions. To clear differences, a tedious manual survey is usually required to collect the information on actual building conditions. By using an RGB-D camera, this study combines the point cloud SALM and multi-feature pose estimation of image sequences to collect 3D point clouds of on-site conditions and realize indoor positioning. Thus, change areas can be determined by comparing point clouds with the corresponding BIM model and then introduced to the proposed point cloud recognition process, which can greatly improve the automation of information collation and reduce the labor and time cost for BIM-based facility management. Because of this, the project designs a two-year plan, in which the research scope comprises indoor positioning, multi-feature image pose estimation, on-site point cloud registration, object point cloud recognition, object reconstruction, and quality assessment. Moreover, the proposed methods will be evaluated by simulated and real datasets to verify their feasibility and effectiveness. This project introduces a low-cost and highly automated mechanism for BIM model renewal, which can facilitate the work of space configuration and equipment inventory in facility management. Notably, considering the topic of 3D point cloud recognition attracts research attention in recent years, this study would render significant academic contributions and assist in the development and implementation of BIM-based facility management in the industry.

Decision Support and Self-positioning of an Intelligent Sewer Inspection Platform

智慧下水道巡檢平台之決策支援與空間定位

Considering the impact of short-time and heavy rainfall, regular maintenance of drainage systems including internal structure inspection of drainage pipelines and pipeline dredging plays an essential role in disaster prevention and reduction. In contrast to most existing robotic approaches that merely use video for visual examination, the study develops an intelligent pipeline inspection platform, consisted of low-cost components such as optical, infrared, and range imagery as well as a g-sensor, to conduct sewer inspection and self-positioning with learning-based image processing techniques. The collected image sequences are used to detect the internal defects and obstacles within a pipeline and estimate the sectional area occupied by an obstacle. Moreover, the flatness and inclination of a pipeline are analyzed by observing the vertical acceleration and the orientation of the platform. The effectiveness of the proposed method has been validated by preliminary experiments, revealing that the platform not only improves the automation level of sewer inspection but also saves costs in labor and time.

Steel Structure Industry 4.0 via Imaging Control​

​

基於深度影像控制技術進行鋼構廠生產與管理工業4.0技術研發

• 輔助SAW門型電焊機自動焊接 <減少焊接作業人力>

• 內隔板間距自動量測 <提升品質管理自動化程度>   

• 提供低成本，高精度，快速且自動化的解決方案

Gallery

Welcome to DSI Lab.

Get in Touch

1001 University Road, Hsinchu, Taiwan 300, ROC

​​office: Room 323, Engineering Building II

jtychuang[at]nycu.edu.tw

+886 3 5712121 ext: 54916