光學點雲測繪導入品質資訊與互動式策略-以建物為例

Abstract

現有光學點雲測繪軟體與實務操作流程中多忽略品質資訊的記錄與回饋，缺乏明確判斷依據與客觀品質標準。本研究旨在解決光學點雲測繪中品質資訊不足的問題，建構一套品質導向的測繪框架，實現「圖資測繪+品質資訊」的雙重輸出，提升點雲資料的可靠度與應用效能。為闡述核心理念及簡化圖資標的，本研究現階段以建物為主要測繪對象，基於Open3D與PySide6建構使用者友善的互動式測繪平台，導入四項核心策略：(1)系統性誤差檢查：透過外部資料比對確保空間準確性；(2)測繪困難區偵測：建立偵測、篩選與標示策略，協助使用者檢視潛在高風險區域並採取適當測繪決斷；(3)影像倒投影：基於共線方程式建立物空間與像平面幾何關係，分為診斷性與驗證性兩種應用。診斷性倒投影將測繪困難區位置投影回原始影像，協助辨識匹配失敗或資料缺失成因；驗證性倒投影結合誤差傳播分析，除了檢視影像倒投影位置之外，亦計算像點位置標準差及參數誤差影響量分析，提供完整的量測成果精度查驗及誤差分析；(4)重複量測及品質計算：提供點位精度與穩定性指標。 本研究所設計之光學點雲測繪平台功能發揮測繪品質的監控與分析體系，形成多策略整合框架於互動式視覺化平台靈活運作。最重要者，本研究工作補足點雲測繪中長期被忽略的品質拼圖，藉此提供系統性決策支援機制，提升測繪成果的可靠度與應用價值。

Current point cloud surveying and mapping software and operational workflows often neglect quality information recording and feedback, lacking clear judgment criteria and objective quality standards. This study aims to address the problem of insufficient quality information in photogrammetric point cloud surveying and mapping by establishing a quality-oriented surveying and mapping methodological framework that achieves dual output of "mapping results + quality information," enhancing the reliability and application efficiency of point cloud data. For the simplicity and elaborating the effort on the key issues, the research focuses on buildings as the primary mapping objects and constructs a user-friendly interactive measurement platform based on Open3D and PySide6, implementing four core strategies: (1) Checking systematic error: ensuring spatial accuracy through comparison with external data; (2) Detecting difficult area of surveying and mapping: establishing detection, filtering, and marking strategies to assist users in examining potentially high-risk areas, thus making appropriate decision ; (3) Image reprojection: establishing geometric relationships between object space and image space based on collinearity equations supporting two applications. Diagnostic reprojection projects surveying and mapping difficulty areas back to original images to help identify causes of matching failures or data deficiencies; verificative reprojection, on the other hand, combines error propagation analysis to calculate image point coordinates as well as their standard deviations and offer the error analysis for a complete accuracy assessment; (4) Repeated measurement and quality calculation: providing point accuracy and stability indicators. The research outcomes include a quality monitoring and analysis system, forming a multi-strategy integrated framework, and developing an interactive visualization platform. Last but not least, this study fills the long-neglected quality puzzle in point cloud surveying and mapping, providing systematic decision support mechanisms and enhancing the reliability and application value of surveying and mapping results.