Impact of various Vegetation Indices on Mango orchard mapping using Object-Based Image Analysis

Abstract

Mango farming is an important part of the Indian agriculture economy. Mapping of mango orchards is essential for monitoring mango plantations as well as its yield assessment. Object-based Image Analysis (OBIA) is a powerful image classification method which uses spatial and spectral information for image classification. This study assesses the impact of three vegetation indices; NDVI (Normalised Difference Vegetation Index), ReNDVI (Red Edge Normalised Difference Vegetation Index) and LSWI(Land Surface Water Index) on the accuracy of classification using object-based image analysis using Sentinel - 2 data. A temporal profile was generated to select the best possible dates for classification based on the maximum and minimum values of the index. LSWI gave the highest overall accuracy of the classification (89%) followed by ReNDVI (87%) and NDVI (86%).The study found that LSWI and ReNDVI have the potential for better mapping of Mango orchards and can be explored further to generate accurate Mango orchard maps.