Determination of textural indices of guava fruit using discriminate analysis by impact force

Abstract

This research developed an effective and nondestructive method for inspecting the texture of guava fruit and highly sensitive indices of quality using an impact pendulum and discriminate analysis. Impact parameters calculated from the force-time curve, amplitudes, and spectra were used to reflect the variation in the texture of fruit during storage. Statistical discriminate analysis was conducted to determine the impact parameters and highly sensitive classification indices from combinations of parameters. Test results indicate that the two classifications by the indices, obtained from parameter sets of TA (force-time and amplitude spectrum) and AI (amplitude spectrum and imaginary part spectrum), were both on average over 80% consistent with the change in the maturity of the fruit, which was significantly more than the strength of the relationship with any single impact parameter, which yielded a classification accuracy of under 70%.