Deteksi Kualitas Telur Ayam Ras Berdasarkan Standar Nasional Indonesia Menggunakan Metode EfficientDet

Abstract

Chicken eggs are among the most popular food ingredients in Indonesia. This is due to their relatively affordable price and their widespread use as complementary foods and essential ingredients in food production. The distribution of eggs by sellers is typically based on size. However, irresponsible sellers often exploit this system by selling other types of eggs as consumption eggs, which poses a risk for culinary business operators who require large quantities of high-quality eggs. To address this, SNI 3926:2023 can be applied to classify the quality of chicken eggs as consumable products. Egg quality is categorized into three grades: Grade I, Grade II, and Grade III, which are assessed both internally and externally. In terms of external quality, one of the determining factors is shell thickness, which can be inferred from the shell's color. However, since the color differences among chicken eggs are often subtle, culinary business operators need considerable time to select high-quality eggs. To overcome this challenge, a detection approach is required to assess the quality of chicken eggs efficiently. In this study, the EfficientDet method is utilized to detect the quality of chicken eggs based on shell color, classifying them into Grade I, Grade II, and Grade III. This approach is implemented in an Android system. Testing results demonstrate that the model, utilizing the EfficientDet algorithm, successfully detects the quality of chicken eggs based on shell color with an accuracy of 96.66%.