Advancing Mango Ripeness Assessment: A Comprehensive Study Integrating VNIR Spectroscopy and SIMCA Modelling for ‘Dashehari’ Cultivar

Abstract

This study addresses the challenges associated with assessing mango ripeness, particularly in the Dashehari cultivar, a popular mid-season mango in northern India. Farmers faced many problems during harvesting season. Traditional ripeness assessment methods are deemed inaccurate and time-consuming, necessitating the development of non-destructive techniques. The research focuses on the application of Visible and Near-Infrared (VNIR) spectroscopy, coupled with chemical models, to create a versatile tool for predicting Soluble Solids Content (SSC) in thin-skinned fruits with similar physicochemical characteristics. The investigation extends to the effectiveness of VNIR spectroscopy in combination with classification models for mango identification and ripening stage prediction. The chosen wavelength regions, guided by preprocessing techniques and Principal Component Analysis (PCA), demonstrate distinct clustering among unripe, half ripe, and fully ripe mangoes, particularly in the 670-850 nm range. The Soft Independent Modelling by Class Analogy (SIMCA) model, incorporating PCA, achieves remarkable classification accuracy rates of 100%, 96.66%, and 93.33% for unripe, half ripe, and fully ripe fruits, respectively, within the 670-850 nm wavelength region. In the context of the Dashehari mango, known for its green skin even when fully ripe, the study provides valuable insights into precise ripeness assessment. The proposed approach holds significance for the mango industry, aiding in quality assurance and post-harvest strategies for marketing, transportation, and storage. The combination of VNIR spectroscopy and SIMCA modelling emerges as a promising solution, offering advantages in terms of accuracy, efficiency, and reduced post-harvest losses.