Local-Export Quality Classification Device for Multiple Unripe Carabao Mango using Single Shot Detector Algorithm

• Published in: 2023 11th International Symposium on Digital Forensics and Security (ISDFS) pp. 1 - 6
• Main Authors: Rosales, Menchie, Marie Culpable, Regiena, Cayanan, Christian, Forcadela, Gino, Mae Boctil, Gilian, Cecilia Venal, Maria
• Format: Conference Proceeding
• Language: English
• Published: IEEE 11.05.2023
• Subjects: carabao mango; deep learning; Detectors; Image color analysis; local and export quality; Manuals; Production; Real-time systems; Streaming media; Universal Serial Bus; visual inspection
• DOI: 10.1109/ISDFS58141.2023.10131810

Mangoes are among the most extensively exploited fruits for food, juice, flavor, fragrance, and color in the Philippines. Agriculture is growing fast paced in terms of modern technology, but the main issue is that the mango industry is still using traditional ways of classifying the carabao mango local and export product. There are different studies classifying the quality of mango with different deep learning algorithm, but these existing papers doesn't have a test about how fast it is compared to the manual processing and will classify multiple mangoes using Single Shot Detector algorithm. This paper has a device that can classify in real-time video processing, gather data, and display information on the harvested unripe carabao mangoes. It has a feature that could capture and classify mangoes using a USB camera connected to the device. Also, the system will display the class or grade of mangoes whether for local and export with at least 15 mangoes at a time. The machine or the device results in a much faster of 37.4 seconds than the manual process. This means that it will save 60-70% of the time for the farm in identifying the export quality unripe mangoes with 97.5 percent accuracy over the four testing cases.