利用BIM與機器視覺技術輔助鋼筋施工檢驗

傳統鋼筋檢驗是以人工現場目測檢查為主，施工人員與品管人員常因判斷標準不一致引起彼此爭議影響檢驗效率，甚至於影響施工品質。機器視覺運用於成品檢驗在工業上已行之有年，在土木營造業的相關研究也越來越多，本研究為改善鋼筋人工檢驗作業，應用了建築資訊模型(Building Information Modeling, BIM)結合機器視覺技術來輔助鋼筋尺寸與間距的現場檢測工作，希望在鋼筋施工檢驗上以科學方式量得的客觀數據作為判斷標準，讓品管人員與施工人員間有一把共同的尺，免除爭議同時提供建築資訊上更完善的品管數據紀錄。本研究利用ImageJ影像分析軟體將BIM模型鋼筋圖檔與實體鋼筋影像進行影像加強前處理、角隅檢測及二值化等作業，取得鋼筋之間的角隅點數據後再建置一套鋼筋尺寸與間距量測演算法，最後比對模型鋼筋與現場鋼筋的間距及尺寸差異達到自動鋼筋檢測的目的。

關鍵字

機器視覺；角隅檢測；鋼筋檢驗；鋼筋間距；鋼筋尺寸

並列摘要

In practice, construction inspectors rely primarily on visual inspection for rebar acceptance. Controversies arise when the inspector and the rebar installer have inconsistent judgment standards, resulting in low inspection efficiency and sometimes poor construction quality. In recent years, machine vision technology has been used for quality control in many different industry sectors, and some examples have also been reported in construction. In this study, we combine building information models (BIM) with machine vision technology to improve the inspection process of rebar installation. We develop a scientific inspection mechanism aiming at providing an objective standard for both inspectors and installers. In this study, a commercial image analysis software, ImageJ, is used for image pre-processing, including image enhancement, corner detection, and image binarization. The rebar corner data obtained from pre-processing are used to calculate rebar size and spacing. Finally, we compared the size and spacing between the BIM model and the actual rebar installation image and an objective judgement can be made. With this automated mechanism, a better inspection process and quality data keeping procedure can be achieved.