direkt zum Inhalt springen

direkt zum Hauptnavigationsmenü

Sie sind hier

TU Berlin

Inhalt des Dokuments


Object detection of aerial image using mask-region convolutional neural network (mask R-CNN)
Zitatschlüssel Musyarofah.2020
Autor Musyarofah and Valentina Schmidt and Martin Kada
Seiten 1–11
Jahr 2020
DOI 10.1088/1755-1315/500/1/012090
Journal IOP Conference Series: Earth and Environmental Science
Jahrgang 500
Verlag IOP Publishing
Zusammenfassung The most fundamental task in remote sensing data processing and analysis is object detection. It plays an important role in classification and very useful for various applications such as forestry, urban planning, agriculture, land use and land cover mapping, etc. However, it has many challenges to find an appropriate method due to many variations in the appearance of the object in image. The object may have occlusion, illumination, viewpoint variation, shadow, etc. Many object detection method has been researched and developed. Recently, the development of various machine learning-based methods for object detection has been increasing. Among of them are methods based on artificial neural network, deep learning and its derivatives. In this research, object detection method of aerial image by using mask-region convolutional neural network (mask-R CNN) is developed. The result shows that this method gives a significant accuracy by increasing the image training and epoch time.
Link zur Publikation Download Bibtex Eintrag

Zusatzinformationen / Extras


Schnellnavigation zur Seite über Nummerneingabe