Detection of coffee fruits on tree branches using computer vision
ABSTRACT Coffee farmers do not have efficient tools to have sufficient and reliable information on the maturation stage of coffee fruits before harvest. In this study, we propose a computer vision system to detect and classify the Coffea arabica (L.) on tree branches in three classes: unripe (green)...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Universidade de São Paulo
2022-09-01
|
| Series: | Scientia Agricola |
| Subjects: | |
| Online Access: | http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0103-90162023000100103&tlng=en |
| _version_ | 1811262197583577088 |
|---|---|
| author | Helizani Couto Bazame José Paulo Molin Daniel Althoff Maurício Martello |
| author_facet | Helizani Couto Bazame José Paulo Molin Daniel Althoff Maurício Martello |
| author_sort | Helizani Couto Bazame |
| collection | DOAJ |
| description | ABSTRACT Coffee farmers do not have efficient tools to have sufficient and reliable information on the maturation stage of coffee fruits before harvest. In this study, we propose a computer vision system to detect and classify the Coffea arabica (L.) on tree branches in three classes: unripe (green), ripe (cherry), and overripe (dry). Based on deep learning algorithms, the computer vision model YOLO (You Only Look Once), was trained on 387 images taken from coffee branches using a smartphone. The YOLOv3 and YOLOv4, and their smaller versions (tiny), were assessed for fruit detection. The YOLOv4 and YOLOv4-tiny showed better performance when compared to YOLOv3, especially when smaller network sizes are considered. The mean average precision (mAP) for a network size of 800 × 800 pixels was equal to 81 %, 79 %, 78 %, and 77 % for YOLOv4, YOLOv4-tiny, YOLOv3, and YOLOv3-tiny, respectively. Despite the similar performance, the YOLOv4 feature extractor was more robust when images had greater object densities and for the detection of unripe fruits, which are generally more difficult to detect due to the color similarity to leaves in the background, partial occlusion by leaves and fruits, and lighting effects. This study shows the potential of computer vision systems based on deep learning to guide the decision-making of coffee farmers in more objective ways. |
| first_indexed | 2024-04-12T19:20:54Z |
| format | Article |
| id | doaj.art-c60cfaba34ed488b8e7867e294173889 |
| institution | Directory Open Access Journal |
| issn | 1678-992X |
| language | English |
| last_indexed | 2024-04-12T19:20:54Z |
| publishDate | 2022-09-01 |
| publisher | Universidade de São Paulo |
| record_format | Article |
| series | Scientia Agricola |
| spelling | doaj.art-c60cfaba34ed488b8e7867e2941738892022-12-22T03:19:37ZengUniversidade de São PauloScientia Agricola1678-992X2022-09-018010.1590/1678-992x-2022-0064Detection of coffee fruits on tree branches using computer visionHelizani Couto Bazamehttps://orcid.org/0000-0001-5685-1099José Paulo Molinhttps://orcid.org/0000-0001-7250-3780Daniel Althoffhttps://orcid.org/0000-0001-5390-575XMaurício Martellohttps://orcid.org/0000-0003-3251-915XABSTRACT Coffee farmers do not have efficient tools to have sufficient and reliable information on the maturation stage of coffee fruits before harvest. In this study, we propose a computer vision system to detect and classify the Coffea arabica (L.) on tree branches in three classes: unripe (green), ripe (cherry), and overripe (dry). Based on deep learning algorithms, the computer vision model YOLO (You Only Look Once), was trained on 387 images taken from coffee branches using a smartphone. The YOLOv3 and YOLOv4, and their smaller versions (tiny), were assessed for fruit detection. The YOLOv4 and YOLOv4-tiny showed better performance when compared to YOLOv3, especially when smaller network sizes are considered. The mean average precision (mAP) for a network size of 800 × 800 pixels was equal to 81 %, 79 %, 78 %, and 77 % for YOLOv4, YOLOv4-tiny, YOLOv3, and YOLOv3-tiny, respectively. Despite the similar performance, the YOLOv4 feature extractor was more robust when images had greater object densities and for the detection of unripe fruits, which are generally more difficult to detect due to the color similarity to leaves in the background, partial occlusion by leaves and fruits, and lighting effects. This study shows the potential of computer vision systems based on deep learning to guide the decision-making of coffee farmers in more objective ways.http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0103-90162023000100103&tlng=enYOLOprecision agriculturehigh-quality coffee |
| spellingShingle | Helizani Couto Bazame José Paulo Molin Daniel Althoff Maurício Martello Detection of coffee fruits on tree branches using computer vision Scientia Agricola YOLO precision agriculture high-quality coffee |
| title | Detection of coffee fruits on tree branches using computer vision |
| title_full | Detection of coffee fruits on tree branches using computer vision |
| title_fullStr | Detection of coffee fruits on tree branches using computer vision |
| title_full_unstemmed | Detection of coffee fruits on tree branches using computer vision |
| title_short | Detection of coffee fruits on tree branches using computer vision |
| title_sort | detection of coffee fruits on tree branches using computer vision |
| topic | YOLO precision agriculture high-quality coffee |
| url | http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0103-90162023000100103&tlng=en |
| work_keys_str_mv | AT helizanicoutobazame detectionofcoffeefruitsontreebranchesusingcomputervision AT josepaulomolin detectionofcoffeefruitsontreebranchesusingcomputervision AT danielalthoff detectionofcoffeefruitsontreebranchesusingcomputervision AT mauriciomartello detectionofcoffeefruitsontreebranchesusingcomputervision |