Objects classification in google earth images by using fuzzy clustering algorithm and neural networks

Dissertant

al-Rashid, Sura Zaki Naji

University

University of Technology

Faculty

-

Department

Computer Sciences Department

University Country

Iraq

Degree

Master

Degree Date

2006

English Abstract

The data of Google Earth program can provide the object Information of the earth's surface accurately.

Now the commercial satellites can provide the satellite images with the High resolution.

With those high spatial resolution images, shape features and texture features and ether's of the ground objects are very clear.

Therefore, this thesis presents a method to design programming system by using hybrid techniques (Fuzzy Clustering Algorithm and Neural Network Algorithm) and three types from Color Spaces (RGB, YUV, and HSI) to classify these images into the objects with minimum acceptable error depending on their features.

This search is concerned with classifying seven kinds of objects (urban, forests, grassland, streets, water, soil, commercial cities).

Accordingly, firstly we read the image from the Google Earth program with RGB color space (hen we convert this imago another color space is YUV to be used in the segmentation stage only Y band one byte to find the clusters center exact values instead of the three values to clusters center with RGB color space, In the segmentation stage, we use Fuzzy C-Means (FCM) algorithm to segment the image to number of the clusters similar of the attributes.

After this, we extract two types of features for each cluster which are visual features including (Pattern, Shape, Texture, Shadow, Association), and statistical features represented by spectrum features that include (Intensity, Hue, Saturation).

In the last stage, we appoint feed forward neural network from Type error Back-Propagation neural network to determine the class that each feature vector belongs to.

Main Subjects

Information Technology and Computer Science

Topics

• Earth(Planet)

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-305673