A graph generation pipeline for critical infrastructures based on heuristics, images and depth data
This work addresses the need for affordable and accessible digital twins in critical infrastructure management, though it is incremental as it builds on existing photogrammetry and deep learning techniques.
The authors tackled the problem of creating virtual models for critical infrastructures by developing a cost-effective graph generation pipeline using RGB images and depth data from stereo cameras, which produced graphs close to ground truth for two hydraulic systems.
Virtual representations of physical critical infrastructures, such as water or energy plants, are used for simulations and digital twins to ensure resilience and continuity of their services. These models usually require 3D point clouds from laser scanners that are expensive to acquire and require specialist knowledge to use. In this article, we present a graph generation pipeline based on photogrammetry. The pipeline detects relevant objects and predicts their relation using RGB images and depth data generated by a stereo camera. This more cost-effective approach uses deep learning for object detection and instance segmentation of the objects, and employs user-defined heuristics or rules to infer their relations. Results of two hydraulic systems show that this strategy can produce graphs close to the ground truth while its flexibility allows the method to be tailored to specific applications and its transparency qualifies it to be used in the high stakes decision-making that is required for critical infrastructures.