Mango (Mangifera indica cv. Sein Ta Lone) ripeness level prediction using color and textural features of combined reflectance-fluorescence images

Abstract

Sein Ta lone mango of different maturity level images has been obtained using reflectance and fluorescence imaging systems. It has been found that fluorescence images show interesting patterns in correlation with the accumulation of bluish fluorescence compounds in the lenticel spots on the mango surface. Color and textural features of both reflectance and fluorescence images have been evaluated to develop a ripeness prediction model. The results show that combining color features of reflectance image and textural features could increase the R2 of the Partial Least Square Regression (PLSR) model up to 0.97 for Brix prediction and 0.99 for pH prediction with Root Mean Square Error (RMSE) of 0.5 for both. These results show the potential of the combined reflectance-fluorescence imaging system for mango ripeness assessment.

Keywords

• Machine vision
• Spectroscopy
• Textural features
• Haralick
• Ripeness