Adrasso 2020

Offentlig prosjektsammendrag

The project goal is to develop and demonstrate an intelligent, autonomous drone for inspections of ships and off-shore vessels. The drone will fly by itself in a cargo or ballast tank, track its position, and have intelligence to spot rust, cracks and inspect coating condition, measure steel thickness, and compare with historical data to see the development of rust and cracks over time. With this the industry will improve ship safety, personnel safety, inspection quality and efficiency, as well as environmental impact. Status per July 2020 is that we have developed and are developing the following: 1. an AI computer vision algorithm that can automatically detect cracks and corrosion in color images 2. tested technical feasibility of hyperspectral imaging to detect and classify corrosion and paint types 3. a drone suited for inspection i GPS-denied environments, and equipped with a number of capabilities that are required for inspection drones in harsh environments, for example indoor navigation, localization and collision avoidance, and taking thickness measurements based on giving only high-level commands that capture the intent of the drone pilot 4. a hyperspectral camera design (Optical requirements) for a low-end (low weight and low price) hyper-spectral camera for integration on the drone; it is cheaper, ligher and smaller than the current high-end hyperspectral cameras 5. collected and prepared a large amount of images containing cracks and corrosion that are used to train the AI computer vision 6. produced over 100 test panels for the hyper-spectral image analysis