Geometry-based automated recognition of objects on satellite images of sub-meter resolution
Скачать файл:
URI (для ссылок/цитирований):
https://link.springer.com/chapter/10.1007%2F978-3-030-51974-2_36https://elib.sfu-kras.ru/handle/2311/142344
Автор:
Mozgovoy, D. K.
Kapulin, D. V.
Svinarenko, D. N.
Yamskikh, T. N.
Chikizov, A. A.
Tsarev, R. Y.
Коллективный автор:
Институт космических и информационных технологий
Кафедра разговорного иностранного языка
Кафедра информатики
Дата:
2020-07Журнал:
Advances in Intelligent Systems and ComputingКвартиль журнала в Scopus:
Q3Квартиль журнала в Web of Science:
без квартиляБиблиографическое описание:
Mozgovoy, D. K. Geometry-based automated recognition of objects on satellite images of sub-meter resolution [Текст] / D. K. Mozgovoy, D. V. Kapulin, D. N. Svinarenko, T. N. Yamskikh, A. A. Chikizov, R. Y. Tsarev // Advances in Intelligent Systems and Computing. — 2020. — № 1226. — С. 371-379Аннотация:
The paper considers an algorithm for automated classification of mobile small size objects on multispectral satellite images of submeter spatial resolution using radiometric and geometric features. It ensures recognizing the desired classes of objects with high accuracy regardless of their orientation in the image. The geometric features of the objects classified in the binary image included the area of the object, the lengths of the principal and auxiliary axes of inertia, the eccentricity of the ellipse with the main moments of inertia, the area of a convex polygon described near the object, the equivalent diameter of a circle with the same area, and the convexity coefficient.