Academic Report Notice of Sergey Abrameyko: Object and change detection in remote sensing images

发布者:王健发布时间:2023-07-04浏览次数:27

Speaker: Academician Sergey Abrameyko

Title: Object and change detection in remote sensing images

Time: 9:30amJuly 5th, 2023 (Wednesday)

Location: Report hall of  Yifu

Abstract:  

In systems of remote sensing image monitoring the quality of automatic classification of objects is largely dependent on the results of object detection. The real images subject to distortions caused by the presence of shadows, changing lighting conditions and scenes of their own imaging equipment noise. The presentation will show the developed technologies for remote sensing image processing, object and change detection and map updating. Next, we will introduce the effective tool developed for object detection, change detection, and updating digital maps through remote sensing images. We will demonstrate and explain the two application scenarios of this tool for detecting discrete objects (buildings, ships), slender objects (roads), and detecting changes in images taken at different times. To detect objects at UAV (super-high resolution - 0.1–0.3 m) images, we developed an approach based on the proposed deep neural network architecture integrating the Faster R-CNN with the Feature Pyramid Network module. We proposed an algorithm to detect shadows on the base of definition of its structure and features on urban or built-up satellite images by using neural networks. The results of object detection will be shown. Finally, the practical experience in creating a comprehensive processing system for remote sensing and map images, as well as the future application scenarios of the developed technology, will be discussed.

Personal Introduction:

    Sergey Abrameyko, an academician of the Belarusian National Academy of Sciences, the Belarusian National Academy of Engineering, the Russian Academy of Natural Sciences, the Academia Europaea, the Asia Pacific Association of Artificial Intelligence, and the International Pattern Recognition Association, is currently a professor of the Belarusian State University. Academician Sergey is a recipient of the Belarusian National Award for Best Young Scientist, the Belarusian National Award (the highest national science award), and the Belarusian National Academy of Sciences Award for Best Scientist. Academician Sergey has published 67 works and over 300 academic papers in international journals and conferences. Academician Sergey has served as the editor in chief of international journals such as Informatics, Vesti Nacionalnoi Academii Nauk Belarusi, series: physics, mathematics, information, and deputy editor in chief of Vestnik Vyshei Shkoly.


[Editor: Tong Gao]