The effective detection of objects in remote sensing images is of great research importance, so recent years have seen a significant progress in deep learning techniques in this field. However, despite much valuable research being conducted, many challenges still remain. A lot of research projects focus on detecting objects of a single category (class), while correctly detecting objects of different categories is much harder. The recognition of small and overlapping objects is often very problematic. The highest valued classifiers are universal ones that help accurately detect objects of various categories. This research project compared the efficiency of detecting objects of various categories, such as airports, helicopters, planes, fuel tanks and warships, using various modern neural network architectures in the public remote-sensing dataset for geospatial object detection (RSD-GOD). The results presented in this paper are better than the results of detecting objects of the same categories in the RSD-GOD dataset produced by previous studies.
Authors
- Aleksander Madajczak,
- dr hab. Marcin Ciecholewski link open in new tab
Additional information
- DOI
- Digital Object Identifier link open in new tab 10.1109/igarss53475.2024.10640694
- Category
- Aktywność konferencyjna
- Type
- publikacja w wydawnictwie zbiorowym recenzowanym (także w materiałach konferencyjnych)
- Language
- angielski
- Publication year
- 2024
