enhancement of the neural network performance to classify remote sensing data by wavelet transform. | ||
| Journal of Education and Science | ||
| Article 19, Volume 26, Issue 1, March 2013, Pages 42-56 PDF (0 K) | ||
| DOI: 10.33899/edusj.2013.89460 | ||
| Author | ||
| Lubna Thanoon Alkahla* | ||
| Abstract | ||
| ABSTRACT During the last period, many technique had been developed to increase the enhancement of the neural network performance and increase it's ability in the classification of remote sensing data. Research include a study of wavelet transform effect to improve the neural network classification which is applied on satellite images and extract the land features of the classified area with high quality and sensitivity. The study involve preprocessing on satellite images represented by spectral packages of scence of red sea taken by Thematic mapper, depending on the multiresolution analysis that provided by wavelet transform, then classify the enhancement images by the use of unsupervised neural network (ART1). Depending on preprocessing images and by calculating the classification accuracy for each class, the results showed notified enhancement of the primary data involved in the classification and increased the accuracy of resulting classes in comparison with initial images prior to management. | ||
| Keywords | ||
| neural network; Remote Sensing Data; wavelet transform | ||
|
Statistics Article View: 323 PDF Download: 324 |
||