Advances in Geospatial Technology for Early Detection of Mango Leaf Diseases

Abstract

Mango leaves are susceptible to pests and pathogens including anthracnose, powdery mildew, diebacks, bacterial infection which cause serious loss of mango production, quality and sustainability. Recent developments in geospatial technology have provided new opportunities for the early, accurate, and large-scale detection of disease. This article offers a comprehensive review of methods, datasets and tools for monitoring mango leaf diseases through remote sensing and geoscience. High-resolution UAV images, multispectral and hyperspectral satellite data (e.g. Sentinel-3) along with field level spectral information serve as the basic sources of data for disease detection. There is a need for preprocessing steps such as resizing, denoising and normalization to enhance the reliability of models. The feature extraction with the help of vegetation indices, Band Ratios and texture metric GLCM increases the discrimination capacity between healthy and infected leaf. SVM, Random Forest, CNNs, and hybrid models of machine learning and deep learning have demonstrated high performance in classification when trained on a well-processed dataset. A key advance presented in this review is the development of portable spectrometers, which provide very accurate leaf-level reflectance profiles and become the crucial ground truth for calibrating geospatial disease detection models. The combined use of spectrometer data with UAV and satellite imagery significantly enhances disease identification sensitivity, especially at early stages of disease infection when the visual symptoms are minimal. In general, this review confirms that combining geospatial technologies, spectroscopy, and smart classification frameworks will provide an efficient, accurate, and scalable solution for mango disease monitoring to support precision horticulture and sustainable orchard management.