Sayed Muchallil, Silvia Roza, Muhammad Hafez Al-Assad, Yudi Candra, Maya Fitria, Rahmad Dawood
In 2022, Indonesia produced roughly 3.203.775 metric tons of pineapples. The sweetness of pineapples is an important factor in determining their selling price. Typically, a pineapple’s sweetness is determined manually, which is time-consuming and prone to human error, resulting in a lower selling price and pineapple waste due to unsold fruit. To address this issue, we developed a deep learning-based algorithm for identifying the sweetness of pineapples based on the hypothesis that a pineapple’s sweetness is closely related to its color and maturity. To train the algorithm, we took 872 images of freshly harvested Smooth Cayenne pineapples and measured their sweetness with a Brix measuring device. We use two algorithms, EffisientNetB4 and RestNet-50, both with a learning rate of 0.000001, to differentiate between sweet and not-so-sweet pineapple. Both algorithms detected pineapple sweetness with the same accuracy of 84.09%. However, RestNet50 had a greater loss than EffisintNetB4. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd 2023.
Department of Electrical and Computer Engineering, Universitas Syiah Kuala, Aceh, Banda Aceh, 23111, Indonesia; Telematics Research Center, Universitas Syiah Kuala, Aceh, Banda Aceh, 23111, Indonesia
Research at a Glance
Register to unlock
Topics & SDG Alignment
Register to unlock
Collaboration
Register to unlock
Author Profile (Selected)
Register to unlock
References Overview
Register to unlock
Journal & Source
Register to unlock
Metadata & Integrity
Register to unlock