Grey Oyster Mushroom Classification toward a Smart Mushroom Grading System for Agricultural Factory

A visual evaluation of vegetables and fruits helps classify tasks in handling the agricultural product quality. For this, the grey oyster mushroom requires attention for having a smart grading system for helping the agricultural farmers to ease their job in grading them before transporting mushrooms to a supermarket. Mainly in Malaysia, Federal Agricultural Marketing Authority (FAMA) is an agricultural organization that proposed a categorical type for grey oyster mushrooms; premier, grades 1, 2, and 3 based on mushroom's size, color, pattern, and damage condition. Manual eye observation is still conducted in this traditional classification for grey oyster mushrooms. Even though most research works are conducted with the inter-class task for a bigger group of mushroom classification, the intra-class mushroom study is still limited and needs attention, which compares the available image to a smaller level or class. In this paper, intra-class classification, image processing, and deep learning are utilized, with an introduction of grey oyster mushroom classification using CNN as our classification engine is proposed. Here, it is necessary to evaluate the grading system traditionally conducted manually by farmers based on the FAMA guidance. The study also aims to design a grey oyster mushroom classification network model. The VGG16 pre-trained network model will be studied and utilized in this study. This study's results should introduce systematic training, testing, and evaluation for grey oyster mushroom classification for agricultural activities by evaluating the range of possible design solutions utilizing machine learning by a new arrangement of mushroom dataset images. Thus, the research performed a comparative and reliable classification accuracy, although the study is still mid-way to accomplishing the full research target, ready with a smart mushroom grading system for the agricultural factory.