A Novel Multi-Index Approach For Crop Health Analysis Using PCA-Derived Spectral Weights

For precision agriculture, resource optimization, and sustainable farm management, prompt and correct crop health assessment is vital. Conventional vegetation indexes like NDVI and EVI offer insightful observations, yet they usually fall short of portraying the whole spectrum of physiological and structural conditions seen in several crops and habitats. By means of Principal Component Analysis (PCA), this study introduces a novel Multi-Vegetation Health Index (MVHI) integrating five complementary spectral indices—NDVI, EVI, SAVI, GNDVI, and NDRE—to derive objective statistical weights. Thirteen scenes from June–December 2024 from Sentinel-2A surface reflectance imagery were preprocessed using cloud filtering, spatial point extraction, and temporal index computation at field sites in Mahalaxmikheda village, Maharashtra. MVHI was built from the first principal component, which accounted for 82.30% of total spectral variation, hence confirming great information compression and index strength. Accuracy assessment using one-way ANOVA shown significant discriminatory ability across independently defined crop health groupings (F = 211.50, p < 0.001), with a large effect size (η² = 0.522). Comparative ANOVA results demonstrated that MVHI offers a balanced alternative to highly sensitive individual indices such NDRE and GNDVI. Spatial mapping of MVHI revealed clear gradients of crop vigor and stress, hence supporting its applicability for field-scale monitoring and agronomic decision-making. MVHI turns out to be a trustworthy, data-driven and whole crop health indicator fit for operational remote sensing and scientific study purposes.