Satellite Data Integration for Citrus Mapping Diyala City, Iraq

Abstract

Diyala is one of Iraq's largest citrus producer areas, characterized by many orange trees and sour and sweet lemon trees embraced by palm orchards. The absence of outstanding quality and spectral remotely sensed information makes distinguishing between distinct fruit tree species with similar spectral and spatial characteristics difficult, particularly for broadleaf evergreen trees. This work provides a unique decision technique for mapping the geographical distribution of citrus trees using spectral reflectance characteristics taken from Landsat 9 band spectral indices data. The spectrum reflectance of citrus trees was calculated using a seasonal series of data (2018-2023) of spectral indicators such as the Normal Difference Vegetation Index (NDVI), Soil Adjusted Vegetation Index (SAVI), Natural Chlorophyll Pigment Index (NPCRI), and Normalized Difference Moisture Index (NDMI). The citrus plants exhibited values of reflectance that varied between 0.4 and 0.6, following the results obtained. The research region was identified using a random forest method that combined the spatial resolution of citrus trees with Google Earth images. The findings indicated that the classification model could differentiate citrus plants and generate information about distribution at a total accuracy (OA) of 97% of the respondents and a kappa statistic of 0.91.