An improved pistachio detection approach using YOLO-v8 Deep Learning Models
Pistachios are an agricultural product widely used in the food industry. It is very important that pistachios are presented to the consumer in good quality on time. At the same time, whether the shells of pistachios are open or closed is an important criterion from a commercial industrial point of v...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
EDP Sciences
2024-01-01
|
| Series: | BIO Web of Conferences |
| Subjects: | |
| Online Access: | https://www.bio-conferences.org/articles/bioconf/pdf/2024/04/bioconf_i-craft2024_01013.pdf |
| _version_ | 1797353017218957312 |
|---|---|
| author | Gökalp Çınarer Mübarek Mazhar Çakır |
| author_facet | Gökalp Çınarer Mübarek Mazhar Çakır |
| author_sort | Gökalp Çınarer |
| collection | DOAJ |
| description | Pistachios are an agricultural product widely used in the food industry. It is very important that pistachios are presented to the consumer in good quality on time. At the same time, whether the shells of pistachios are open or closed is an important criterion from a commercial industrial point of view. Pistachios with their shells open have a high unsaturated fat content, a high maturity level and an expensive market value. In this study, the open or closed status of pistachios was determined by using Artificial Intelligence-based deep learning models. For pistachio detection, 423 image data belonging to the Pesteh dataset were classified using models of the Yolov8 algorithm, which detects objects using convolutional neural networks. The data set is divided into 80% training, 10% validation and 10% testing. The performances of the models were evaluated with precision, recall, F1 and mAP score metrics. The highest test mAP value of the Yolov8 algorithm, which was run with image data consisting of pistachios, was obtained with the Yolov8-m model with 94.8%. The Yolov8-m model achieved a very successful result with 49.6 MB weight size, 11.0 ms inference time value and 0.33 hours training time value. In addition, the model's fast classification performance and small file size facilitate its applicability in the industrial field. The results show that the classification and detection of open and closed shell pistachios has been successfully carried out with Yolo models. |
| first_indexed | 2024-03-08T13:24:08Z |
| format | Article |
| id | doaj.art-f415dfbfe24b42a69b704faf3c4f30db |
| institution | Directory Open Access Journal |
| issn | 2117-4458 |
| language | English |
| last_indexed | 2024-03-08T13:24:08Z |
| publishDate | 2024-01-01 |
| publisher | EDP Sciences |
| record_format | Article |
| series | BIO Web of Conferences |
| spelling | doaj.art-f415dfbfe24b42a69b704faf3c4f30db2024-01-17T15:01:19ZengEDP SciencesBIO Web of Conferences2117-44582024-01-01850101310.1051/bioconf/20248501013bioconf_i-craft2024_01013An improved pistachio detection approach using YOLO-v8 Deep Learning ModelsGökalp Çınarer0Mübarek Mazhar Çakır1Yozgat Bozok University, Department of Computer EngineeringYozgat Bozok University, Department of Mechatronics EngineeringPistachios are an agricultural product widely used in the food industry. It is very important that pistachios are presented to the consumer in good quality on time. At the same time, whether the shells of pistachios are open or closed is an important criterion from a commercial industrial point of view. Pistachios with their shells open have a high unsaturated fat content, a high maturity level and an expensive market value. In this study, the open or closed status of pistachios was determined by using Artificial Intelligence-based deep learning models. For pistachio detection, 423 image data belonging to the Pesteh dataset were classified using models of the Yolov8 algorithm, which detects objects using convolutional neural networks. The data set is divided into 80% training, 10% validation and 10% testing. The performances of the models were evaluated with precision, recall, F1 and mAP score metrics. The highest test mAP value of the Yolov8 algorithm, which was run with image data consisting of pistachios, was obtained with the Yolov8-m model with 94.8%. The Yolov8-m model achieved a very successful result with 49.6 MB weight size, 11.0 ms inference time value and 0.33 hours training time value. In addition, the model's fast classification performance and small file size facilitate its applicability in the industrial field. The results show that the classification and detection of open and closed shell pistachios has been successfully carried out with Yolo models.https://www.bio-conferences.org/articles/bioconf/pdf/2024/04/bioconf_i-craft2024_01013.pdfyolov8pistachios detectiondeep learningartificial intelligence |
| spellingShingle | Gökalp Çınarer Mübarek Mazhar Çakır An improved pistachio detection approach using YOLO-v8 Deep Learning Models BIO Web of Conferences yolov8 pistachios detection deep learning artificial intelligence |
| title | An improved pistachio detection approach using YOLO-v8 Deep Learning Models |
| title_full | An improved pistachio detection approach using YOLO-v8 Deep Learning Models |
| title_fullStr | An improved pistachio detection approach using YOLO-v8 Deep Learning Models |
| title_full_unstemmed | An improved pistachio detection approach using YOLO-v8 Deep Learning Models |
| title_short | An improved pistachio detection approach using YOLO-v8 Deep Learning Models |
| title_sort | improved pistachio detection approach using yolo v8 deep learning models |
| topic | yolov8 pistachios detection deep learning artificial intelligence |
| url | https://www.bio-conferences.org/articles/bioconf/pdf/2024/04/bioconf_i-craft2024_01013.pdf |
| work_keys_str_mv | AT gokalpcınarer animprovedpistachiodetectionapproachusingyolov8deeplearningmodels AT mubarekmazharcakır animprovedpistachiodetectionapproachusingyolov8deeplearningmodels AT gokalpcınarer improvedpistachiodetectionapproachusingyolov8deeplearningmodels AT mubarekmazharcakır improvedpistachiodetectionapproachusingyolov8deeplearningmodels |