Survey Paper on Determining the Rice Seeds Quality Using Convolutional Neural Network

Plant nurseries and farmers depend on seed checking to guarantee seed quality for seedling growth. Traditionally, skilled inspectors sift samples by hand, but there are a number of difficulties, including expense, accuracy, and volume.Accuracy and speed were the primary requirements for raising agricultural output. Rice seed quality classification is one area of study that can benefit from the application of machine learning, a branch of artificial intelligence. Training, validation, testing, and dataset gathering make up a machine learning system's pipeline. Creating a model starts with gathering information on the physical attributes of rice seeds, such as their color and shape. Two thousand photos from two categories—superior seeds and non-superior seeds—make up the dataset. On Google Collaboratory notebooks, training and validation were carried out utilizing the Convolutional Neural Network (CNN) method and the cross-validation concept. When modeling from a dataset, the ratio of train data to validation data is 80:20. The model's output is a Deep Convolutional Neural Network (Deep CNN) that can categorize digital images of rice seeds based on the outcomes of data calls that are entered into the system. With a precision value of 93% and a recall of 95%, the system can distinguish between superior and non-superior seeds based on the analysis of the experiment's 30 test data.