การเพมประสทธภาพความแมนยำในการจำแนกความสกของทเรยนดวยการตดภาพพนหลงแบบไฮบรด
RMUTP Research Journal Sciences and Technology · 01.01.2025
DOI: 10.14456/jrmutp.2025.15 ↗Kurz gefasst
Fachartikel in der Zeitschrift RMUTP Research Journal Sciences and Technology, erschienen am 01.01.2025. Verfasst von รติพร จันทร์กลั่น.
Abstract (Originalsprache)
Durian is an economically important fruit in Thailand. It is widely consumed, both domestically and internationally. As a result, durians play a significant role in stimulating the economy, both nationally and internationally. Among the various varieties, the Monthong durian stands out as the most popular. Durian can be consumed in either fresh or processed forms. For fresh consumption or purchasing durian for processing purposes, the ripeness level of the durian is a critical factor to consider. However, without skilled classification, rip or unrip durians may lead to the acquisition of that does not meet their own needs. Currently, artificial intelligence techniques have been applied to fruit maturity classification to help consumers choose suitable fruits by using fruit images, sometimes having a background image attached to them. When processing these images, it is necessary to remove the background to improve the efficiency of data classification. In some cases, using existing knowledge in image processing to remove the background of an image using a single background removal algorithm may not be sufficient for certain types of data. Therefore, this research presents the enhancement of deep learning efficiency in classifying the ripeness of durians with hybrid background removal, edge detection with Sobel algorithms and Grab Cut algorithms for clipping image background then create model with 4 deep learning algorithms to compare the classification performance between using the original image data and image background clipping, the result shown that the proposed techniques can improve the classification performances with average the 4 algorithms at 8%.
Schlagwörter
Zitieren
APA
จันทร์กลั่น, ร. (2025). การเพมประสทธภาพความแมนยำในการจำแนกความสกของทเรยนดวยการตดภาพพนหลงแบบไฮบรด. RMUTP Research Journal Sciences and Technology. https://doi.org/10.14456/jrmutp.2025.15
Vancouver
จันทร์กลั่น ร. การเพมประสทธภาพความแมนยำในการจำแนกความสกของทเรยนดวยการตดภาพพนหลงแบบไฮบรด. RMUTP Research Journal Sciences and Technology. 2025. doi:10.14456/jrmutp.2025.15
BibTeX
@article{2025,
author = {จันทร์กลั่น, รติพร},
title = {การเพมประสทธภาพความแมนยำในการจำแนกความสกของทเรยนดวยการตดภาพพนหลงแบบไฮบรด},
journal = {RMUTP Research Journal Sciences and Technology},
year = {2025},
doi = {10.14456/jrmutp.2025.15}
}
EUVIMED