It is the overarching goal of DIG-IT! to connect computer science developers and ecologists, particularly with respect to the application of deep convolutional neural networks (DCNNs). By doing this DIG-IT! aims to make the opportunities that the digitization offers available for ecology and to answer urgent questions that are relevant to our society. Regarding the research questions, methods, research organization, science policy and utilization of the results, here some summarizing perspectives:
1. The connection of basic ecological research with automatic image aggregation and classification allows for a quantum leaps in the respective disciplines and a wide variety of cases.
2. The currently available ‘cutting edge’ methods in the area of image evaluation (DCNNs), experimental plant ecology (in-situ root monitoring), Paläoökologie (non-pollen palynomorphic analysis, high resolution poll-density dating), landscape ecology (high resolution wood anatomy), nature conservation (wildlife population monitoring) and biomathematics (graph theoretical trees) will be connected (1.) and advanced.
3. The applications, algorithms and tool-kits developed in DIG-IT! will be made available to other scientists in MV. This ‘catalysing’ and multiplicative effect of DIG-IT! will connect and advance existing research projects in MV.
4. DIG-IT! will create an international visible research cluster by combining competences and advancing methodologies (2.), expanding staff and intensifying cooperations (3.), which will significantly contribute to our understanding of the environment and its reaction to climate and land-use change.
5. Through the cooperation with existing national projects (WETSCAPES, RESPONSE) the potential to raise further large projects on a national and international level will be advanced and ideally the method portfolio will be enlarged.
6. Via its qualification program and knowledge transfer, DIG-IT! will create highly qualified jobs and attractive employment opportunities in MV. DIG-IT! will foster ‘digitally competent’ ecologists which will be able to tackle future questions with a state-of-the art set of methods.
7. The developed methods and procedures for the automatic image analysis will be transferred to the market in several ways. Mainly this will be digital tools for ecological service providers like the automatic recognition of animal and plant species, habitats and biotopes based on a complex combination of various traits (e.g. relevant for the permission to build on- and off-shore wind power facilities).
