Main Page
Deanship
The Dean
Dean's Word
Curriculum Vitae
Contact the Dean
Vision and Mission
Organizational Structure
Vice- Deanship
Vice- Dean
KAU Graduate Studies
Research Services & Courses
Research Services Unit
Important Research for Society
Deanship's Services
FAQs
Research
Staff Directory
Files
Favorite Websites
Deanship Access Map
Graduate Studies Awards
Deanship's Staff
Staff Directory
Files
Researches
Contact us
عربي
English
About
Admission
Academic
Research and Innovations
University Life
E-Services
Search
Deanship of Graduate Studies
Document Details
Document Type
:
Thesis
Document Title
:
Pattern Classification for Land Cover Analysis in Satellite Images
التصنيف النمطي لتحليل الغطاء الأرضي في صور الأقمار الصناعية
Subject
:
Pattern Classification for Land Cover Analysis in Satellite Images
Document Language
:
Arabic
Abstract
:
Land cover classification is extremely important for land use planning and management. Pattern classification is a wide research area that can classify and analyze the Earth’s surface to study the natural resources and to understand many environmental phenomena. This research aimed to apply pattern classification methods to classify satellite images into several land cover types such as water, plants, rocks, and desserts. This was done by presenting and implementing a proposed classification system. Satellite images of two different study areas from Jeddah in Saudi Arabia were acquired by Landsat 7 ETM+. This research presented the application of pattern recognition algorithm, the Support Vector Machines (SVM) that is a machine learning algorithm based on statistical learning theory. SVM carried out the process of classification, and gave promising results in land cover classification problem. Ranklet transform is a recent image processing technique proposed and applied in this research as a feature extraction method. Ranklet transform is a multi-resolution and orientation selective approach similar to that of the wavelet transform; in addition, it is non-parametric that deals with the ranks of the pixels rather than with their gray-level intensity values. Ranklet coefficients as classification features could solve successfully the classification problem and gave superior results in classification performances Furthermore, with an accurate choice of the SVM parameters and of the resolutions at which the multi-resolution ranklet transform is performed, ranklet-based images achieved an improvement in accuracy (91%) over the pixel–based (84%) and wavelet–based (86%) images. Performance evaluation was the final step in the proposed classification system, which performed in term of accuracy.
Supervisor
:
Dr.
Thesis Type
:
Master Thesis
Publishing Year
:
1429 AH
2008 AD
Added Date
:
Monday, May 4, 2009
Researchers
Researcher Name (Arabic)
Researcher Name (English)
Researcher Type
Dr Grade
Email
سحر حسين جمبي
Jambi, Sahar Hussein
Researcher
Master
Files
File Name
Type
Description
19910.pdf
pdf
Back To Researches Page