An Industrial Dataset for Scene Acquisitions and Functional Schematics Alignment

TL;DR

This paper introduces IRIS-v2, a comprehensive dataset to support further research, aiming at reducing the time required for functional schematics alignment by combining segmentation and graph matching.

Abstract

Aligning functional schematics with 2D and 3D scene acquisitions is crucial for building digital twins, especially for old industrial facilities that lack native digital models. Current manual alignment using images and LiDAR data does not scale due to tediousness and complexity of industrial sites. Inconsistencies between schematics and reality, and the scarcity of public industrial datasets, make the problem both challenging and underexplored. This paper introduces IRIS-v2, a comprehensive dataset to support further research. It includes images, point clouds, 2D annotated boxes and segmentation masks, a CAD model, 3D pipe routing information, and the P&ID (Piping and Instrumentation Diagram). The alignment is experimented on a practical case study, aiming at reducing the time required for this task by combining segmentation and graph matching.

Figures (7)

• Figure 1: This paper introduces IRIS-v2 (violet), a new dataset based on IRIS armangeon2026iris (blue) that provided a dense point cloud and a CAD model in a room exceeding $530 m^2$, and was used for point visibility estimation armangeon2025iris-vis. In addition to IRIS data, IRIS-v2 introduces 300 spherical images, functional schematics with $\mathord{\sim}500$ pieces of equipment, pipe routing, 6000 annotated boxes and 47000 segmentation masks. These data are used for the problem of scene acquisitions and functional schematics alignment.
• Figure 2: Data modalities of the IRIS-v2 dataset. (a) point cloud, (b) CAD model reconstructed close to the points, (c) 300 spherical images and the P&ID, (d) pipelines routing, (e) annotated 2D masks, (f) annotated 2D boxes.
• Figure 3: Example of alignment between scene acquisitions and functional schematics. The problem consists in locating pieces of equipment from the functional schematics in the 3D scene. Since information on distances is not available in the schematics, the matching is done based on structure relationships between objects. The filter and vibration isolators cannot be segmented as they are hidden by heating isolation.
• Figure 4: 2D detection and 3D Segmentation results using our fine-tuned Grounding DINO model liu2023grounding and SAM kirillov2023segment for 2D segmentation and standard 2D-to-3D masks projection and fusion for 3D segmentation.
• Figure 5: Scene graph construction steps. P nodes are pipe elements and A, B and C nodes denote equipment types.