Mining of Very High Resolution Satellite Images Using Hierarchical Graph Models
Summary:
The global coverage and the capability to capture information about both very large areas and small local details have made satellite imagery a vital data source for many applications involving topics such as environment, agriculture, disaster management, security, health, energy, natural resource management, urban planning and mapping. The constant increase in the amount of available data as well as the urgent need for the extraction of useful information from raw data necessitated automatic or semi-automatic analysis of such images.
The main goal of this project is to develop novel methods for the automatic analysis of very high spatial resolution images obtained from the new generation sensors. The methodology aims to achieve a multi-scale semantic abstraction of the image content by using statistical and structural properties of the image data. The proposed work involves the design and development of a hierarchical model that aims to detect compound structures in images by exploiting the spectral and spatial properties of simpler image primitives.
People:
- Faculty
- Selim Aksoy (Principal investigator)
- Graduate Students
- Hüseyin Gökhan Akçay (Ph.D. student)
- Çağlar Arı (Ph.D. student)
- Ahsen Ikbal Yergök(M.S. student)
Duration:
April 2010 - October 2012
Sponsor:
TUBITAK - Scientific and Technological Research Council of Turkey (Grant no: 109E193) |
Budget:
184,954 TL (~US$123,000)