Nowadays, even when adherence to project schedule and budget are the most critical performance metrics among project owners, still 53% and 66% of typical construction projects exhibit schedule delays and cost overruns, respectively. Intending to contribute to a more efficient construction progress monitoring, this thesis proposes a method to detect some of the most common temporary objects classes in a laser scanner point cloud of a construction site. These objects can provide a precise estimate of the current construction progress. For this purpose, this thesis focuses on the detection of cranes, scaffolds, and formwork. The method involves computer vision and machine learning techniques to detect vertical instances of the selected object classes. The proposed workflow begins with the automatic downsampling and rotation of the point cloud. Subsequently, the target objects are detected using a combination of several techniques: image processing over vertical projections, finding patterns in 3D detected contours and performing checks over specifically generated vertical cross-sections. A deep learning algorithm was leveraged to classify these cross-sections for the purpose of formwork detection. The method was applied on three real point clouds of construction sites to assess its accuracy. The results reveal that the method achieves average rates above 88% for precision and recall. Moreover, the technique also achieved outstanding computational time performance. This demonstrates the capability of the method to support the automatic segmentation of point clouds of construction sites. Further development can be done to increase the precision and automation of the technique.     «

Nowadays, even when adherence to project schedule and budget are the most critical performance metrics among project owners, still 53% and 66% of typical construction projects exhibit schedule delays and cost overruns, respectively. Intending to contribute to a more efficient construction progress monitoring, this thesis proposes a method to detect some of the most common temporary objects classes in a laser scanner point cloud of a construction site. These objects can provide a precise estimate...     »