#TITLE_ALTERNATIVE#
Information about objects in satellite image depends on wavelength value used. Different objects have diferent reflection respond especially for radiated and reflected wavelength. Collection of same object will have same characteristics (features) at its reflected waves.<p>Backpropagation, whi...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/10777 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Information about objects in satellite image depends on wavelength value used. Different objects have diferent reflection respond especially for radiated and reflected wavelength. Collection of same object will have same characteristics (features) at its reflected waves.<p>Backpropagation, which is one of feed forward neural network with supervised learning, is used as objects classifier in satellite image. Training or learning on backpropagation network works based on gradient descent. Learning method that used for this classification is based on Levenberg-Marquardt error minimization.<p>Features that already extracted from satellite image is used as input for backpropagation network. The use of co-ocurence matrix before image feature extraction can improve quality of classification images. In this paper, water and not water is result of classification category. <br />
|
|---|