KAJIAN KEMAMPUAN JARINGAN SYARAF TIRUAN BERBASIS CITRA ALOS DALAM IDENTIFIKASI LAHAN KRITIS (Studi Kasus Kecamatan Dlingo dan Sekitarnya)
| Main Authors: | , NURSIDA ARIF, , Drs. Projo Danoedoro, M.Sc.,PhD. |
|---|---|
| Format: | Thesis NonPeerReviewed |
| Terbitan: |
[Yogyakarta] : Universitas Gadjah Mada
, 2012
|
| Subjects: | |
| Online Access: |
https://repository.ugm.ac.id/97882/ http://etd.ugm.ac.id/index.php?mod=penelitian_detail&sub=PenelitianDetail&act=view&typ=html&buku_id=52952 |
Daftar Isi:
- Identification of critical land is generally done using the scoring method by overlaying maps of determinants (variables) of critical land. Non-scoring methods such as Artificial Neural Networks is rarely used. This research is expected to be an alternative reference method in the determination of critical land that often uses spatial data. The purpose of this study was to determine the accuracy of the identification of critical land using Artificial Neural Networks by comparing the results of classification using spectral data and non-spectral data in the identification of critical land, and determine the effect of changing parameters of Artificial Neural Networks on the accuracy of the identification of critical land (iteration, the hidden layer, momentum, learning rate and RMS error). The research methodology consists of several stages of data collection, correction of image radiometry and geometry, field orientation, selection of the training areas, execution of the classification results using artificial neural network methods, as well as determining the accuracy of the test samples. The critical land parameters used are vegetation cover, slope, and depth of solume and soil texture. The sampling method used in this study was stratified random sampling method. The results of the research are in the form of critical lands map derived using neural network classification. The highest accuracy is obtained in a simulation using 7 channels by combining spectral and non spectral data that is 83.33%, occurred in the parameter with 1 (one) hidden layer