Development of Object Oriented Bare Surface Feature Extraction Algorithm for Desertification Early Warning using Nigeria Sat-2 High Resolution Imagery Data

Taiwo Qudus; Shuaibu Ade; Mohammed ALIYU Modibbo; Umar Wali; Mustapha Abdulrahman Lawal; Ismail Zahraddeen Yakubu1

1

Publication Date: 2024/01/03

Abstract: Federal Government of Nigeria in 1999 embarked on some laudable steps at pioneering the development of Space Technology in Africa which eventually led to the launching of Nigeriasat-1 on the 27th of September 2003 by National Space Research and Development Agency (NASRDA), Abuja, Nigeria. However, it is pertinent to know that since the launching of Nigeriasat-2 satellite, algorithm for the extraction of feature classes from its image has not been developed. Therefore, this study is not only exploring the techniques for the extraction of feature classes from Nigeriasat-2 VHRI but with a view to develop algorithm that will be more accessible than foreign based algorithms like eCognition to average Earth Observation Scientists. Secondly, Desertification is land degradation that occurs in arid, semi-arid and dry sub-humid areas. Over the years, desertification and drought are two related disasters largely contributing to high rate of famine, especially in the Northern part of Nigeria. This research has been conducted in order to monitor desertification extent and severity over northern Nigeria using developed algorithm from object oriented normalized Bare Surface index (nBSI). The Global Positioning System (GPS) was used specifically for picking sample points during ground sampling and reconnaissance survey. Other image processing software such as Erdas Imagine, Idrisi Selva,Ilwis and ENVI were used during various image processing procedure. Visual basics and Java software were used for algorithm and final software development. Six criteria are used to evaluate the performance of a classification method: accuracy, reproducibility, robustness, ability to fully utilize the data's information content, uniform applicability, and objectiveness.. This method proves valuable to local and regional governments, educational and research institutions, and private businesses as a result of its simplicity, low cost, and integration with a range of software products.

Keywords: Object Oriented Bare Surface, Feature Extraction Algorithm, Desertification, Early Warning, NigeriaSat-2 and High Resolution Imagery Data.

DOI: https://doi.org/10.5281/zenodo.10453284

PDF: https://ijirst.demo4.arinfotech.co/assets/upload/files/IJISRT23DEC098.pdf

REFERENCES

No References Available