SISTEM PENDETEKSI SOLUBLE SOLID CONTENTS (SSC) PADA BAGIAN MEMAR (BRUISES) BUAH BERBASIS CITRA VIS-NIR

Bahasa Indonesia

Abstract

The time prediction system and the distribution of SSC content in bruises are determined based on the length of time of storage using Vis-NIR images at a wavelength of 400-1000 nm using a system without having to damage the object. So that the information obtained can not only distinguish the bruised area, but also provide information on the depth of the bruise and SSC content in the area around the bruise. The Vis-NIR image system used consists of a set of devices, including a workbench, a slider, two halogen light sources (150W) and a Vis-NIR image camera connected to a PC via Camera Link. The system software consists of measurement of reflectance image profiles, feature extraction, feature selection on spectral and spatial data, PLSR models of SSC content, and PLSR models of bruising depth. Partial Least Square Regression (PLSR) model is used to develop prediction models for all wavelength spectral data. The PLSR model was used to predict the value of SSC content and the depth of the bruise. The predicted results were compared with the results of laboratory tests of SSC content obtained using a refractometer and the depth of the bruises obtained was measured by an expert. From the results of the performance of the prediction model, the RMSE value from the testing data is 0.06 and the correlation coefficient from the testing data is 0.99.