DEVELOPING A GEODATABASE FOR EFFICIENT UAV-BASED AUTOMATIC CONTAINER CRANE INSPECTION

Inspection of container cranes is an important task to maintain the all-day operation in harbours. In practice, manually carrying out the inspection process is not the best solution due to the complexity, time consumption, and cost. The manual inspection often is performed by specialists who capture images of areas of interest and then check and analyse these images visually. Once a critical area is spotted, industrial climbers are hired to check the part in-situ. The visual pre-inspection should be reliable and accurate, because the in-situ inspection is expensive, but to miss defected areas might lead to a failure of the crane. From this point of view, we came up with the main idea of the joint research project ABC-Inspekt which embodies an exploratory investigation in providing an automatic inspection of container cranes using the technology of Unmanned Aerial Vehicles (UAV). Images, which are captured systematically and regularly, are managed in a database to facilitate access. The access is done by an operator, but also by an image analysis approach which automatically identifies possible defects and stores them back into the database. In this paper, we introduce the database schema, the web frontend for data storage, access, and viewing, and show intermediate results for the data processing workflow.