The aim of the project "ABC-Inspect" is to achieve a significant optimization and more efficient execution of container crane inspections with simultaneous cost savings. This contributes to an increase in equipment reliability, the handling performance of the port terminals and the smooth inflow and outflow of goods on the seaside.
In order to maintain the uninterrupted and all-day operation of container bridges (24 hours / 365 days) in a seaport, it is extremely important to carry out a qualified inspection. Meanwhile, drones/multicopters are used to obtain extensive image material of neuralgic construction areas of the container bridges. Changes in the surface of the construction (colour irritation, surface curvature, rusting) must be detected at an early stage, otherwise consequential damage or even fractures of the container bridges may occur. The visual evaluation of the image material captured by the drones is currently carried out manually with the help of qualified specialists. Analyses are carried out and decisions are made on the basis of personal assessment, experience and the employee's current physical condition. The results of an investigation can vary here; there are also risks of not recognizing certain suspicious facts.
A self-learning, automatic image recognition system based on an AI approach is to be used to evaluate the UAV-based images and automatically compare changes in the same areas and regions of the container gantry cranes over a longer period of time. The images, evaluation results and annotations are stored in a powerful database, which is intended to improve the documentation of suspected defects, improve maintenance processes and provide security with regard to liability issues.
Image Processing | Database |
---|---|
Dr.-Ing. Mehdi Maboudi | Dr.-Ing. Ahmed Alamouri |
M.Sc. Vanessa De Arriba López | M.Sc. M. Shafi Bajauri |