In this project I developed various deep learning models for semantic segmentation that can be used by biologists for biomonitoring. Semantic segmentation for biomonitoring offers several advantages. Large amounts of data can be processed automatically, which saves a lot of time that experts need. At the same time, quantitative analyses can be carried out, allowing a much deeper insight into different plant communities. Neighborhood analyses of individual species and their dynamic changes can be modeled to gain a better understanding of how individual organisms interact in a community.