Eggplant Quality Classification Using Backpropagation Algorithm Based on Color and Texture Features

Abstract

Manual quality assessment of eggplant is often inconsistent, takes a long time, and is prone to errors due to worker fatigue. This research aims to develop an automated system based on digital image processing to assess eggplant quality efficiently and accurately. The stages begin with image capture using a mobile phone device designed to ensure stable lighting and uniform background. The acquired image is then processed through segmentation using the Otsu thresholding method as well as morphological operations to separate the main object from the background. Color and texture features are extracted through Gray-Level Co-occurrence Matrix (GLCM) analysis and RGB, HSV, and LAB color spaces. Training data amounting to 90% of the total dataset was used to train an artificial neural network-based classification model with a backpropagation algorithm, while the remaining 10% was used for testing. Experimental results showed that the combination of LAB, RGB, HSV, and texture features gave the best results, with a testing accuracy of 86%, recall of 85%, and precision of 92%. This model is very effective in detecting poor quality eggplants with 100% accuracy. This system can support the application of technology in the horticultural sector.