Quality assessment of chilies (Capsicum annuum L.) by using a smartphone camera

Abstract

This study aims to determine the image parameters that can the chili quality and arrange the best Artificial Neural Networks (ANN) architecture in order to classify the quality into several groups. The image was taken and analyzed by means of image processing and ANN. Samples of chilies were divided into 2 groups, 75% for training data and 25% for network testing. The sample was placed in a black box and the image was captured using smartphone camera for both sides. The sample image was converted to binary to get the color value. Value color of image processing results using Matlab were used as input parameters for color group identification of ANN. Subsequently, color used as variable for classification using 'if then' logic. The results showed that the color elements of the image that could be the quality identity of chili were red, green, a∗, and intensity. The ANN input layer that consisted of 4 input cells, 9 hidden layer cells, and 1 output layer cell could classify chilies into 4 color groups i.e.green, tinge of red, red and dark red with an accuracy of 90.43%.