Cell Phone Image-Based Persian Rice Detection and Classification Using Deep Learning Techniques

Rice stands as a foundational agricultural product and staple food, instrumental in feeding more than half of the global population. It is a significant source of sustenance for approximately 3.5 billion individuals worldwide and represents a crucial element of food security, with an annual production surpassing 500 million tons. Beyond its role as a dietary staple, rice cultivation is a vital economic activity, offering substantial income for countless farmers across various regions. The emphasis on sophisticated and accurate methodologies for rice quality and classification has become increasingly prominent. This urgency is driven by the potential to enhance market acceptability, minimize rejection rates, and elevate the economic gains for producers through reliable quality assurance practices [1]. In the realm of agricultural quality assessment, traditional methods often depend on manual inspection based on visual appearance and smell, which, despite their widespread use, suffer from limitations in speed, accuracy, and reliability, particularly for those without extensive experience. Recent advancements in technology have paved the way for the application of data mining and machine learning techniques, marking a significant leap in enhancing the efficiency and precision of rice classification processes. These innovative approaches utilize detailed feature extraction from images, analyzing color, shape, and textural characteristics to differentiate rice varieties and ascertain their quality with unprecedented accuracy [2, 3, 4, 5, 6]. Sumaryanti et al. present a system designed for the identification of rice varieties using image processing techniques and a LVQ neural network algorithm.