The aim of the AseeDiZi project is to enable predictions to be made about the short and medium-term development of post-mining lakes. This includes evaluating current measures to stabilise water quality (in particular neutralisation through liming) and forecasting future biogeochemical and ecological developments. This forecasting tool is based on digital twins of the post-mining lakes, which in turn require a spatially and temporally high-resolution data set that characterises the water body on a biological, chemical and physical level. This data requirement, which includes both the water body itself and its surroundings, cannot be covered by previous manual surveying techniques and stationary measuring probes.

The project is therefore developing a robotic solution in which a collaborative measurement concept is designed for a drone and a floating robot. Based on known information, the drone is used to collect phenomena that can be detected on the surface and then the robot is used to carry out localised sampling with a lowerable sensor set at depth. The measurement strategy for the floating robot is derived from the drone data. Particular attention will be paid to recording the vegetation in and around the water, which serves as a critical indicator for the development of water quality. The images taken with an underwater camera are analysed using an AI-based classification process to determine the species composition of the aquatic plants. For this purpose, online procedures are being tested that can evaluate the classification quality during image acquisition in order to trigger an adjustment of the image acquisition if necessary.