Comparison between supervised and unsupervised image classification techniques

مقدم أطروحة جامعية

Ibrahim, Muhammad Ahmad Braima

مشرف أطروحة جامعية

al-Sinnari, Adil Muhammad Ahmad

أعضاء اللجنة

al-Junayd, Ammar Babikir Mahmud
Ahmad, Abd al-Rahim al-jazuli Muhammad

الجامعة

جامعة أم درمان الإسلامية

الكلية

كلية العلوم و التقانة

القسم الأكاديمي

قسم نظم المعلومات

دولة الجامعة

السودان

الدرجة العلمية

ماجستير

تاريخ الدرجة العلمية

2011

الملخص الإنجليزي

Image classification has been one on the most vivid research areas in the field of computer vision over the last years.

The main objective of this study is to make comparison between supervised and unsupervised classification to evaluate and estimate what method shall be used to properly classify imagery.

The study was conducted at al-Gazira State which is located in the middle of Sudan within the mudflats area until the Upper Nile, most of the surface of state is a plain flat, with no hills, only a few hills in some parts of East and South.

In order to achieve the objective of this study, image enhancement and classification techniques were applied to satellite image of the study area.

Spectral Landsat 5 TM image which acquired on 28th December, 2010 used here as the reference dataset, and pulled to image to georeferencing it.

Six classes to Land-use Land-cover area in the same supervised and unsupervised classification was done in the image.

The study comes to conclusion that supervised classification show some features like water, urban area, etc more clearly and remote sensing is one of strongest new technologies for monitoring and evaluating and exploring earth recourses to make thematic and geomorphologic map which used in GIS in study area to support the decision making.

التخصصات الرئيسية

تاريخ و جغرافيا

الموضوعات

• السودان
• الاستشعار عن بعد
• نظم المعلومات الجغرافية

عدد الصفحات

74

قائمة المحتويات

• APA
• MLA
• AMA

لغة النص

الإنجليزية

نوع البيانات

رسائل جامعية

رقم السجل

BIM-364393