|RESEARCH - Temporal Change Monitoring|
Within environmental research scale temporal studies are an important to understand biological, ecological and geological processes. However, they subsume the repeated scanning of an area of interest for detecting changes over time which is a difficult and time-consuming task if carried out manually by the scientists (i.e., divers), in addition to depth limitations of this type of exploration. Commercial robots are costly, and require a consequent infrastructure for their deployment, that is typically out of reach to the scientific community. The fundamental goal of the project outlined here is to simplify both one-time or repeated explorations providing researchers with a, easily deployable and operated robotic system able to autonomously rescan a scene based on the data, i.e. the global mosaic image and georeference information, acquired during a first exploration.
In addition to the pure physical rescanning of a scene, we also aim at building a system that allow us or scientists to add sensors to complement the imagery and also detect changes in other environmental parameters (i.e., temperature, turbidity, CTD and other miniaturized sensors). Mosaic comparison includes the identification of corresponding areas in both mosaic images. This is to be done by extending local and global image alignment techniques towards an alignment of full mosaic images and an appropriate handling of large-scale mosaics covering widespread areas of the seafloor. The mosaic alignment, however, only yields the base for change detection which by itself requires the development of suitable techniques for image difference detection. Detected differences may perspectively be further classified automatically due to biologically or geologically relevant criteria.