Detection and maturity index classification for tomato: Deep learning with computer vision-based

Nur Khalidah, Z*., Muhd Akhtar, M.T., Nuraini, A.A.S., Seri ‘Aisyah, H., Siti Hawa, A.R., Mohamad Saiful Nizam, A., Mohd Daniel Hazeq, A.R., Mohd Shukry, H.B. and Muhamad Syahiran Afieff, A.

Abstract

This paper is about the development of the tomato detection and maturity index classification model in a greenhouse using a deep learning technique. In total, two deep learning models were developed where the output of the first model will be the input for the second model. In this study, 2000 tomato image samples were captured for data acquisition. The annotated image samples were used to train the tomato detection model. On the other hand, the labeled image samples were used to train the tomato maturity index classification model. The confidence score of the first model was 0.958 whilst the accuracy of the second model was 92.33%. Lastly, both models were deployed and a dashboard was built where users can monitor the total distribution of tomatoes at one time.

Full text (876KB)