Search

Defects in a construction site is inevitable and these defects must be inspected and reported timely to minimize extract cost due to repair the defects. However, in practice, quality of structural elements is often inspected by a site supervisor with traditional tools (e.g. measuring tapes, total stations or levelling) at only specific locations on the structures. That results the current inspection pipeline is subjective and inefficient. As such, a new approach must develop to support project managers reporting surface defects timely and using a digital tool in project management efficiently. A great achievement of laser scanning platforms allows to acquire three-dimensional (3D) topographic data of structures' surfaces in a construction site quickly and accurately. A current terrestrial laser scanner (TLS) can capture more than a million points per second with a sub-millimetre accuracy. This technology has been gradually implemented in evaluating progress, quality and quantity, and visualisation for construction management. However, a processing point cloud requires intensive labour work because of the complexity of the construction site and massive data points. This paper is to develop the algorithm to automatically access structural components for quality control of the construction project, in which the proposed method is focused to evaluate the flatness of the floors and ceiling. The method starts to decompose the point cloud of the building storey into 2D cells by using a quadtree. Subsequently, a combination of kernel density estimation (KDE) and cell-based segmentation (CbS) to extract the data point affiliated surfaces of the floor and ceiling. Next, edges of the slabs are detected and then are used to generate the reference surface to compute deformations by using both point- and cell-surface methods. The proposed method is tested on 23.9-million-point cloud of a storey of the reinforced concrete building acquired from the TLS. An experimental test shows, Optical and Laser Remote Sensing

Defects in a construction site is inevitable and these defects must be inspected and reported timely to minimize extract cost due to repair the defects. However, in practice, quality of structural elements is often inspected by a site supervisor with traditional tools (e.g. measuring tapes, total stations or levelling) at only specific locations on the structures. That results the current inspection pipeline is subjective and inefficient. As such, a new approach must develop to support project managers reporting surface defects timely and using a digital tool in project management efficiently. A great achievement of laser scanning platforms allows to acquire three-dimensional (3D) topographic data of structures' surfaces in a construction site quickly and accurately. A current terrestrial laser scanner (TLS) can capture more than a million points per second with a sub-millimetre accuracy. This technology has been gradually implemented in evaluating progress, quality and quantity, and visualisation for construction management. However, a processing point cloud requires intensive labour work because of the complexity of the construction site and massive data points. This paper is to develop the algorithm to automatically access structural components for quality control of the construction project, in which the proposed method is focused to evaluate the flatness of the floors and ceiling. The method starts to decompose the point cloud of the building storey into 2D cells by using a quadtree. Subsequently, a combination of kernel density estimation (KDE) and cell-based segmentation (CbS) to extract the data point affiliated surfaces of the floor and ceiling. Next, edges of the slabs are detected and then are used to generate the reference surface to compute deformations by using both point- and cell-surface methods. The proposed method is tested on 23.9-million-point cloud of a storey of the reinforced concrete building acquired from the TLS. An experimental test shows, Optical and Laser Remote Sensing