The field of environmental monitoring is witnessing significant advancements with the application of deep learning techniques. Researchers are leveraging these methods to improve the efficiency and accuracy of monitoring systems, particularly in areas such as sea ice classification, coral reef surveying, and building roof type classification. A notable trend is the development of benchmarks and datasets, such as IceBench and Coralscapes, which facilitate the evaluation and comparison of different models. Another key direction is the exploration of self-supervised learning approaches, including the use of foundation models, which have shown promising results in remote sensing applications. These advancements have the potential to enhance our understanding of environmental systems and inform conservation efforts. Noteworthy papers include the introduction of IceBench, a comprehensive benchmarking framework for sea ice type classification, and the Coralscapes dataset, which enables semantic scene understanding in coral reefs. The proposed unified image-dense annotation generation model, TIDE, also shows great potential for alleviating data scarcity issues in underwater scenes.