SEMI-AUTOMATIC RECOGNITION OF OBJECTS FROM CAD DRAWINGS FOR QUANTITY SURVEYING

Abstract

Many engineers still utilize computer-aided design (CAD) drawings for design and then use those drawings for quantity surveys. CAD is user- friendly and easy to learn, but less powerful than Building Information Modelling (BIM). Both technologies are utilized for Bill of Quantity (BOQ). The manual technique for producing BOQ takes a lot of time and may result in errors or model shortcomings. The current study aims to present an accurate framework for object recognition from CAD drawings and convert them into intelligent drawings through which engineers can automatically count amounts. The work here is classified as supervised classification using the principles of remote sensing to identify objects in CAD drawings. This approach aims to convert engineering work into computerized tasks, as it is necessary for large projects. The proposed framework incorporates the C# programming language with Microsoft Excel and AutoCAD. The framework has been tested on over 200 layouts. Drawing faults are detected and corrected automatically or semiautomatically, yielding precise output. The current work minimizes human mistakes in quantity surveys, increases productivity, and completes project information. The study highlights the importance of using the latest technology to overcome the shortcomings in CAD drawings.