Identifikasi Jenis Buah Salak Menggunakan Algoritma SSD-MobileNet

Abstract

Salak is a tropical fruit that has various varieties with subtle but significant morphological differences. This research aims to identify snake fruit types using the SSD Mobilenet method in digital image analysis. The method developed in this research is able to distinguish three types of salak commonly found in Indonesia, namely pondoh salak, sidempuan salak, and Balinese salak, based on physical characteristics such as skin color, size and shape of the fruit. Tests were carried out using a dataset of salak images that had been collected exclusively, and the results showed that SSD Mobilenet achieved satisfactory accuracy. The overall average accuracy achieved was 91.6%. These findings have potential applications in the agricultural industry for more efficient and accurate management of crop varieties, as well as in the food industry to ensure product authenticity. Object recognition technology such as that developed in this research can be used to increase precision in fruit identification, speed up the sorting and identification process..