Image analysis based on color, shape and texture for rice (Oryza sativa L.) seed ageing evaluation system
Keywords:(artificial neural network, color feature, convolutional neural network, rice-seed ageing, shape feature, texture feature).
AbstractThe objective of this research is to develop a computer system, which can evaluate the rice-seed ageing. The system called “rice seed ageing evaluation system or (RiSAES)”, evaluated the rice (Oryza sativa L.) seed ageing by using an image processing technique. The RiSAES consists of five modules, namely: 1) image acquisition, 2) image pre-processing, 3) feature extraction, 4) seed ageing evaluation and 5) result presentation. The RiSAES employed color, shape and texture features of rice-seed images to evaluate the ageing. The system applied the artificial neural network (ANN) and convolutional neural network (CNN) to perform the evaluation. The system precision rates are 81.29% and 80.89% for ANN and CNN, respectively. The average evaluation time for ANN is 0.0528 seconds/image and for CNN is 1.6006 seconds/image.
Download data is not yet available.