| Author | Azcarraga, Arnulfo P. |
| Call Number | AIT Thesis no. CS-87-1 |
| Subject(s) | Image processing
|
| Note | A thesis submitted in partial fulfillment of the requirements
for the degree of Master of Science, School of Engineering and Technology |
| Publisher | Asian Institute of Technology |
| Abstract | This study provides a theoretical basis for the use of
spatial features in the computer recognition of linear objects
in satellite images. Spatial features based on shape, size, and
texture are given precise definitions. Algorithms for measuring
these features are formulated and are based on the chain code
of t h e li near region an d its me dial axis. The simulation of the
photointerpretation process uses a decision tree classifier
involving the measured features. Provision for the estimation
of the confidence level on the interpretation is included .
The theoretical work is verified for its soundness and
applicability through the implementation of a prototype of the
de signed system. The prototype is written in FORTRAN 77 a nd is
tested using a SPOT satellite image of the Vientiane Region.
The system, running on a n IBM 3083, can recognize linear
objects including rivers, streams, highways, railroads , and
airstrips. |
| Year | 1987 |
| Type | Thesis |
| School | School of Engineering and Technology (SET) |
| Department | Other Field of Studies (No Department) |
| Academic Program/FoS | Computer Science (CS) |
| Chairperson(s) | Sadananda, Ramakoti
|
| Examination Committee(s) | Khang, Bach Hung ;
Galtier, Bertrand ;Kozminski, G.
|
| Scholarship Donor(s) | German Academic Exchange Service (DAAD) |
| Degree | Thesis (M.Sc.) - Asian Institute of Technology, 1987 |