12 Nov, 2020
This meetup will be with Port of Rotterdam and ProRail and how they are using state of the art technology to automate asset management.
Talk 1: Port of Rotterdam
Port of Rotterdam has many physical assets which require frequent inspection. Currently this inspection is carried out by an asset manager, driving around the property, assessing the status of selected assets.
As you might have guessed, Port of Rotterdam is running a pilot to smarten things up. Equipping the inspection vehicle with multiple camera’s, labeling collected frames and training machine learning models to detect, localize and assess the quality of relevant assets. The first use case surface estimation of “Sleuven”, which will be soon be succeeded by several other assets which have already been labeled.
We will show you how the models were selected, trained and optimized, and how Azure is supporting the machine learning process, allowing for data versioning and real-time monitoring of performance.
Speakers: Vicky Vercauteren & Laurens Koppenol
Keywords: MaskR-CNN, Azure Machine Learning, Python
Talk 2: ProRail
ProRail is responsible for the maintenance of more then 7000KM of railway roads throughout The Netherlands, so imagine the impact if one track gets broken! Currently the inspection of all these tracks is done manually.
ProRail together with support from Dataworkz is shifting from manually inspecting the tracks to automatically determine if maintenance is needed by using state of the art Computer Vision and Deep Learning techniques. In this talk they will explain how.
Speakers: Oscar van Hees & Willem Bressers
keywords: Azure, Machine Learning, Python
Talk at Hamburg Data Science meetup.