The Future of Visual Inspection for Civil Infrastructure
Speaker: A. Cristiano I. Malossi – Rüschlikon, SwitzerlandTopic(s): Artificial Intelligence, Machine Learning, Computer Vision, Natural language processing
Abstract
Until recently bridge inspection was exclusively a manual process conducted by reliability engineers. Not only is this dangerous due to the complexity of the structure, but it is also estimated to cost a staggering 50B$ and 2B$ person-hours annually.
The advances in drone technology and its failing costs have recently pushed this laborious process of manual inspection progressively towards automation. Drones can now carry high resolution cameras and sensors. Adding to the equation autonomous flight capabilities and advanced AI analytics, these lead to novel ways to bring impactful and at times lifesaving solutions.
In this talk we provide an overview of our visual inspection technology that can be leveraged to conduct fully autonomous inspections of civil infrastructures, such as bridge piers, airport runways, dams and other large structures.. Our end-2-end solution combines state-of-the-art navigation algorithms that enable drones to scan an asset completely autonomously, with novel AI methods that are specialized primarily for detection and segmentation tasks of very small defects in large areas with high precision. Further, they provide invaluable support to reliability engineers by additionally measuring these defects at millimeter level and assessing their severity relative to their position in the structure, by reconstructing overview images of infrastructure elements.
Throughout the talk we also present a live DEMO of IBM One Click Learning (OCL). IBM OCL a research-industry platform that leverages deep learning and advanced computer vision methods to accelerate and improve accuracy in critical inspection tasks and aims at providing end-to-end support for the full data science process. We show how the platform enables civil engineers to conveniently use machine learning algorithms without any AI expertise. This use case example is illustrative of the major roadblocks for adoption of ML in new application domains: the diversity in user profiles and familiarity with data science among domain practitioners; the variety inavailable hardware infrastructure and computing needs; and the heterogeneity, specificity and unsettled evolution of the usecase landscape. Removing these roadblocks requires designing a platform explicitly geared towards usability goals of broad accessibility on one hand, and extensibility and specialization on the other one. We elaborate on a set of functionalities to support these usability goals and thereby enable the design of a task agnostic end-to-end platform to drive ML adoption and workflow standardization in new dynamic application domains. Finally, we present the IBM OCL platform as a proof-of-concept implementation of these functionalities and validate it in a use case where computer vision models are deployed to aid visual inspection of a bridge.
About this Lecture
Number of Slides: 25-30Duration: 45 minutes
Languages Available: English, Italian
Last Updated:
Request this Lecture
To request this particular lecture, please complete this online form.
Request a Tour
To request a tour with this speaker, please complete this online form.
All requests will be sent to ACM headquarters for review.