Accuracy assessment of 2d and 3d geometric correction models for different topographic regions in Iraq

Dissertant

Muhammad, Isra Hatim

Thesis advisor

Zabun, Abd al-Razzaq Tarish

University

University of Technology

Faculty

-

Department

Department of Building and Construction Engineering

University Country

Iraq

Degree

Master

Degree Date

2011

English Abstract

-In recent decades, Remote sensing data becomes one of the basic information required for mapping and different applications in geomatics.

It is usually necessary to preprocess remotely sensed data and remove geometric distortion associated with the satellite images.

Geometrically corrected imagery can be used to extract accurate distance, area, and direction information.

In this research, different mathematical models in 2D and 3D cases are investigated and comprised in order to assess the accuracy of these models under different conditions of terrain topography.

Three high resolution satellite QuickBird images (panchromatic 0.6 m in spatial resolution) of three different study areas, with respect to their topography (a flat area, a hilly area and a mountain area) in Iraq have been used in this work.

The flat area is chosen in Baghdad in the middle of Iraq while, the hilly and mountain areas are chosen in Irbil in the north of Iraq.

The work procedures are evaluated by using the direct GCPs and check point coordinates measurements, these coordinates are collected using the DGPS, (type Leica GPS SR20).

In this research, the 2D mathematical models which were used, 1st, 2nd order polynomial, and projective transformation model while, the 3D mathematical models used were, 1st, 2nd order 3D polynomial, and direct linear transformation model.

All these methods are applied for each study area and evaluated through the Matlab environment facilities.

The results of six models (2D and 3D models) in the three studied areas show that the TRMSE range in the flat area is between (0.5176-0.6042 pixel) for GCPs, and between (0.6085-0.6494 pixel) for check points while the TRMSE range in the hilly area is between (0.5520-0.6914 pixel) for GCPs, and between (0.6282-0.7528 pixel) for check points.

On the other hand, the TRMSE range in the mountain area is between (0.5250-0.8003 pixel) for GCPs, and between (0.5976- 0.8508 pixel) for check points.

Also this results show that the 2nd order 2D polynomial and 2nd order 3D polynomial are the best models in the flat area while the 2nd order 3D polynomial and DLT models are the best models in the hilly and mountain areas.

Also the results indicate that the accuracy of 2D models is nearly similar to the accuracy of 3D models in the flat area while in the hilly and mountain areas the accuracy of 3D models is better than the accuracy of 2D models because 3D models take into account the effect of relief displacement that occurs in the hilly and mountain areas by including the elevation of GCPs (Z) in its function.

Main Subjects

Information Technology and Computer Science

Topics

• Remote sensing
• Digital electronics

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-305305