Intelligent Image-Based Recognition of Rice Cultivars Using PSO-Optimized ANFIS

Rice is considered to be one of the most significant staple foods on a global scale, particularly within developing countries such as Iran. In these countries, rice cultivars frequently exhibit considerable variation in terms of quality, price, and characteristics. Accurate identification of rice cultivars is imperative for maintaining market transparency, ensuring quality control, and supporting agricultural decision-making. This study proposes a novel methodology for the classification of rice cultivars, which integrates advanced image processing techniques with an Adaptive Neuro-Fuzzy Inference System (ANFIS) that has been optimised by the Particle Swarm Optimisation (PSO) algorithm. A total of three rice types, which are commonly cultivated in northern Iran, were selected for the study: Native Tarom, Hashemi Tarom, and Pakistani rice. The images from these three rice types were processed in standard conditions using Canny edge detection algorithms. The extraction of morphological features was then utilised for the training and testing of the ANFIS-PSO model. The experimental results obtained achieved a classification accuracy of 99.47%, thereby demonstrating the superiority of the proposed method in comparison to traditional techniques with regard to precision and applicability. Moreover, the method employs a cost-effective and readily accessible approach that lends itself to practical applications, including the development of smartphone-based software for real-time rice identification. This study proposes a non-invasive, efficient, and scalable solution to rice authentication and classification challenges, with potential application in agricultural quality assurance and market regulation.