Enhancing target crop discrimination: a novel shadow detection technique for RGB datasets in mixed agricultural environments

Loading...
Thumbnail Image

Date

2025

Journal Title

Journal ISSN

Volume Title

Publisher

Mapping Sciences Institute Australia

Abstract

Shadows pose significant challenges in smallholder farming systems, where mixed cropping is common. This study introduces two novel techniques: the Hue-Intensity-Green-Blue (HIGB) difference method for shadow detection and the Light Intensity Ratio-Based (LIRB) method for shadow compensation. Their performance was tested against the C3 and NSVDI models using five accuracy metrics on RGB imagery. HIGB consistently achieved superior accuracies (77–95%) compared to NSVDI (63–84%) and C3 (69–81%) in five different crop mixtures. Both the models, HIGB and LIRB, provide an integrated, robust solution for shadow detection and compensation in heterogeneous agricultural environments.

Description

Keywords

Cast shadow, Crop discrimination, Glycine max, Hue intensity, RGB imagery

Citation

Mafuratidze, P., Mutanga, O., Masocha, M., Dube, T. and Sibanda, M., 2025. Enhancing target crop discrimination: a novel shadow detection technique for RGB datasets in mixed agricultural environments. Journal of Spatial Science, pp.1-16.