APPLICATIONS OF ELECTRONIC NOSE TO DETERMINE THE MATURITY OF SAPODILLA USING PATTERN RECOGNITION SYSTEM

Abstract

Sapodilla is a tropical fruit plants that have a distinctive taste and like most people. Currently the determination of fruit maturity with nondesktrutif for this type of fruit is not widely practiced. The purpose of this study is to identify the scent of amber at different levels of maturity, as well as assess the effect of dose of calcium carbide (CaC2) the maturity of the fruit. The samples were stored in 30oC incubator with treatments: 0.1%CaC2, 0.2% CaC2, and control. The level of hardness fruit, and dissolved solids were measured every day for 8 days. For comparison, the determination of the quality of fruit maturity is determined by sensory test. Fruit aroma is measured by electronic nose that uses four sensors (TGS222, TGS 825, TGS826, and TGS2602). The data used is a voltage change of electronic nose readings. Furthermore, the data is used as the basis for pattern recognition system on artificial neural networks (ANN). Results that have been trained ANN was then tested to identify the level of maturity based on foreign data sapodilla. Data aroma is also also analyzed using Principle Component Analysis (PCA). The goal is to classify according to the level of maturity sapodilla. In this study obtained results that the analysis of ANN is able to recognize the maturity of the fruit with a level of accuracy of 90.11%. PCA is also able to identify the maturity of sapodilla fruit quite well.