Estimation of mass and volume of freshly harvested Naga king chilli (Capsicum chinense Jacq.) using computer vision: Exploring changes during different storage days

This study explores the intricate relationship between the physical attributes of Naga king chilli and their mass and volume, exploiting this correlation to assess chilli quality and devise innovative post-harvest machinery. Chilli grading plays a pivotal role in post-harvest activities in the pickle industry. This assessment is particularly crucial when dealing with Naga king chilli, as it necessitates size and mass-based categorization to ensure optimal processing standards. A prediction method of the mass and volume of Naga king chilli based on a computer vision system was introduced in this study. The work aimed to predict the mass and volume of Naga king chilli as functions of its physical properties, measured using image processing techniques, using single and multiple variable regression models such as linear, quadratic, rational, and exponential. Various mass and volume models are studied to determine their relationships with physical characteristics such as length, width, perimeter, projected area, and elongation ratio. In a single variable mass and volume models, the length-based rational and quadratic models showed the best fit, with R² values of 0.89 and 0.85 and RMSE values of 0.053 and 0.075, respectively. In addition, for multivariable models, the exponential model for both mass and volume were the most suitable regression model, with R² values of 0.92 and 0.90 and RMSE values of 0.046 and 0.062, respectively. The findings of the study revealed that the multivariable model has a better fit than a single variable. • Measurement of physical properties of Naga King chilli using computer vision. • Linear, and non-linear algorithms for mass and volume estimation of Naga King chilli. • Quadratic, rational and exponential models found best for estimation of both mass and volume of Naga chilli. • Naga King chilli mass and volume had shown the linear relationship. • Helps in designing an accurate grading and packaging system.