CLOI: A Shape Classification Benchmark Dataset for Industrial Facilities

Abstract

Generation of digital models of existing industrial facilities is labor intensive and expensive. The use of state-of-the-art deep learning algorithms can assist to reduce the modelling time and cost. However large databases of labelled, laser-scanned industrial facilities do not exist to date, henceforth training of deep learning models is not possible. Our paper solves this problem by proposing a new benchmark dataset, which consists of five labelled industrial plants. The labelling schema that we followed for the generation of this dataset is based on the frequency of appearance of industrial object types. We labelled the ten most frequent industrial object shapes as identified in previous work. We present CLOI (Channels, L-shapes, circular sections, I-shapes): a richly annotated large-scale repository of shapes represented by labelled point clusters. CLOI has more than 140 million hand labelled points and serves as the foundation for researchers who are interested in automated modelling of industrial assets using deep learning algorithms.

Virtual LiDAR Simulation as a High Performance Computing Challenge: Toward HPC HELIOS++

“…Citation information: DOI 10.1109/ACCESS.2022.3211072 large-scale airborne laser scanning [12]. Laser scanning can also be used in rural and urban environments such as [13] and [14] who used terrestrial laser scanning (TLS) and MLS, respectively, or even in industrial facilities [15].…”

“…It is possible to define a child dropper B b from its parent B a as depicted in (15). This expression can be understood as an evolution-like adaptive algorithm because the child is expected to adapt better than the parent to the work context based on the information transmitted during evolution.…”

"BIM-to-Scan" for Scan-to-BIM: Generating Realistic Synthetic Ground Truth Point Clouds based on Industrial 3D Models

“…For our use case in the industrial construction sector, latest research includes a publication on a man-ually labeled collection of industrial point cloud data (Agapaki et al 2019). Unfortunately, to this date, the data presented in this work is not accessible to the authors, and therefore the domain remains without a publicly accessible ground truth dataset.…”

“…Furthermore, top-down observation of the levels allows for a simple separation of the labeled point cloud into more compact sub-clouds without losing class-relevant data. In comparison to the categories introduced with the CLOI dataset (Agapaki et al 2019), our approach adds the craft-wise separation as a novel perspective. Additionally, we omit the sepa- ).…”

CLOI-NET: Class segmentation of industrial facilities’ point cloud datasets

“…We solve instance segmentation in this paper through (a) using a CLOI-Instance graph connectivity algorithm that segments the point clusters of an object class into instances and (b) boundary segmentation of points that improves step (a). Our method was tested on the CLOI benchmark dataset (Agapaki et al 2019) and segmented instances with 76.25% average precision and 70% average recall per point among all classes. This proved that it is the first to automatically segment industrial point cloud shapes with no prior knowledge other than the class point label and is the bedrock for efficient gDT generation in cluttered industrial point clouds.…”