Botanical vegetables recognition on Raspberry Pi using single shot detector (SSD)

Advancements in computer vision technologies have fueled research interest in automating object detection, particularly in agricultural contexts. Human eyes prone to error during the sorting process when differentiating the various types of botanical vegetables such as bell pepper (capsicum), chili,...

Full description

Bibliographic Details
Main Authors: Iqbal Mortadza, M., Shah Zainudin, M.N., Idris, M.I., Mohd Saad, W.H., Kamarudin, M.R., Napiah, Z.A.F.M., Nizam, Nurul Zarirah, Muhammad, Sufri
Format: Article
Language:English
Published: Penerbit UTHM 2024
Online Access:http://psasir.upm.edu.my/id/eprint/115454/
http://psasir.upm.edu.my/id/eprint/115454/1/115454.pdf
Description
Summary:Advancements in computer vision technologies have fueled research interest in automating object detection, particularly in agricultural contexts. Human eyes prone to error during the sorting process when differentiating the various types of botanical vegetables such as bell pepper (capsicum), chili, tomatoes, etc. Hence, the use an object detection method is believed could categorize this botanical vegetables precisely, allowing farmers to optimize their operations and reduce labor expenses. This study explores the identification of various botanical vegetables types using a Raspberry Pi and the Single Shot Detector (SSD). The proposed approach involves curating an extensive botanical vegetables dataset with detailed annotations to optimize training process. Implementing SSD on the Raspberry Pi capitalizes on its processing power and versatility. Our research demonstrates the system's effectiveness in detecting a wide range of botanical vegetables, including chili, capsicum, tomatoes, and vegetable leaf, achieving an average precision of 89% across diverse environmental conditions. Computational efficiency analysis showcases its real-time vegetable detection capabilities, rendering it suitable for agricultural applications such as automated sorting, inventory management, and quality monitoring.