DocumentCode :
2202841
Title :
GANN Snake for Object Extractions from High Resolution Satellite Imagery
Author :
Theng, Lzu Bee ; Ling, Choo Ai
Author_Institution :
Sch. of Comput. & Design, Swinburne Univ. of Technol. Sarawak Campus Kuching, Kuching
fYear :
2008
fDate :
20-22 Dec. 2008
Firstpage :
1005
Lastpage :
1009
Abstract :
Active contour model, well known as snakes are used to extract objects like land parcels and buildings from high resolution satellite imageries like IKONOS and Quickbird. Object extraction from satellite imagery has more a long history. However, increasing image variation, required level of details and higher resolution imagery acquired, object extractions have to be improved continuously. Evolutionary computing approaches can enhance satellite imagerypsilas object extraction. This paper discusses the prototyping of a genetic algorithm (GA) and neural network (NN) snake. The coefficients for snake energy obtained through GANN are compared with ordinary snake for performance.
Keywords :
edge detection; evolutionary computation; feature extraction; genetic algorithms; geophysical signal processing; image resolution; neural nets; random noise; GANN snake energy; active contour model; evolutionary computing approach; genetic algorithm; high resolution satellite imagery; neural network; object edge; object extraction; random noise; resolution imagery; Active contours; Buildings; Data mining; Dynamic programming; Force control; Image resolution; Neural networks; Satellites; Shape; Urban areas; GANN; Genetic Algorithm; Neural Network; satelite imagery;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Advanced Computer Theory and Engineering, 2008. ICACTE '08. International Conference on
Conference_Location :
Phuket
Print_ISBN :
978-0-7695-3489-3
Type :
conf
DOI :
10.1109/ICACTE.2008.149
Filename :
4737109
Link To Document :
بازگشت