@article{HartKnoblachMoeser, author = {Hart, Lukas and Knoblach, Stefan and M{\"o}ser, Michael}, title = {Automated pipeline reconstruction using deep learning \& instance segmentation}, series = {ISPRS Open Journal of Photogrammetry and Remote Sensing}, volume = {9}, journal = {ISPRS Open Journal of Photogrammetry and Remote Sensing}, publisher = {Elsevier}, doi = {10.1016/j.ophoto.2023.100043}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:863-opus-53620}, pages = {19}, abstract = {BIM is a powerful tool for the construction industry as well as for various other industries, so that its use has increased massively in recent years. Laser scanners are usually used for the measurement, which, in addition to the high acquisition costs, also cause problems on reflective surfaces. The use of photogrammetric techniques for BIM in industrial plants, on the other hand, is less widespread and less automated. CAD software (for point cloud evaluation) contains at best automated reconstruction algorithms for pipes. Fittings, flanges or elbows require a manual reconstruction. We present a method for automated processing of photogrammetric images for modeling pipelines in industrial plants. For this purpose we use instance segmentation and reconstruct the components of the pipeline directly based on the edges of the segmented objects in the images. Hardware costs can be kept low by using photogrammetry instead of laser scanning. Besides the autmatic extraction and reconstruction of pipes, we have also implemented this for elbows and flanges. For object recognition, we fine-tuned different instance segmentation models using our own training data, while also testing various data augmentation techniques. The average precision varies depending on the object type. The best results were achieved with Mask R-CNN. Here, the average precision was about 40\%. The results of the automated reconstruction were examined with regard to the accuracy on a test object in the laboratory. The deviations from the reference geometry were in the range of a few millimeters and were comparable to manual reconstruction. In addition, further tests were carried out with images from a plant. Provided that the objects were correctly and completely recognized, a satisfactory reconstruction is possible with the help of our method.}, language = {en} } @article{HartKoblachMoeser, author = {Hart, Lukas and Koblach, Stefan and M{\"o}ser, Michael}, title = {Automation Strategies for the Photogrammetric Reconstruction of Pipelines}, series = {PFG - Journal of Photogrammetry, Remote Sensing and Geoinformation Science}, journal = {PFG - Journal of Photogrammetry, Remote Sensing and Geoinformation Science}, number = {91}, publisher = {Springer Nature}, doi = {10.1007/s41064-023-00244-0}, pages = {313 -- 334}, abstract = {A responsible use of energy resources is currently more important than ever. For the effective insulation of industrial plants, a three-camera measurement system was, therefore, developed. With this system, the as-built geometry of pipelines can be captured, which is the basis for the production of a precisely fitting and effective insulation. In addition, the digital twin can also be used for Building Information Modelling, e.g. for planning purposes or maintenance work. In contrast to the classical approach of processing the images by calculating a point cloud, the reconstruction is performed directly on the basis of the object edges in the image. For the optimisation of the, initially purely geometrically calculated components, an adjustment approach is used. In addition to the image information, this approach takes into account standardised parameters (such as the diameter) as well as the positional relationships between the components and thus eliminates discontinuities at the transitions. Furthermore, different automation approaches were developed to facilitate the evaluation of the images and the manual object recognition in the images for the user. For straight pipes, the selection of the object edges in one image is sufficient in most cases to calculate the 3D cylinder. Based on the normalised diameter, the missing depth can be derived approximately. Elbows can be localised on the basis of coplanar neighbouring elements. The other elbow parameters can be determined by matching the back projection with the image edges. The same applies to flanges. For merging multiple viewpoints, a transformation approach is used which works with homologous components instead of control points and minimises the orthogonal distances between the component axes in the datasets.}, language = {en} }