OPTIMAL SIZE OF AGRICULTURAL DATASET FOR YOLOV8 TRAINING
DOI:
https://doi.org/10.17770/etr2024vol2.8041Keywords:
artificial intelligence, deep learning, precision farming, YOLOv8Abstract
The smart farming solutions are mainly based on the application of convolutional neural networks for object detection tasks. The number of open datasets is restricted in the agricultural domain. Therefore, it is required to find the answer to the question: how big a dataset must be collected to train a convolutional neural network for object detection tasks? To solve this task, the YOLOv8 framework was selected for the experiment. Three datasets were prepared: MinneApples, PFruitlets640 and mosaic dataset using both previously named datasets. 100 images were selected for testing. Other images were used to create training datasets, which had the size from 100 until 1000 images with step 100 images. Training was repeated 10 times with each size of dataset. The experiment showed that the increase of dataset from 100 to 500 images provides an accuracy growth up to 15.48% mAP@0.5, but from 600 to 1000 images - only 2.98% mAP@0.5. This study experimentally proves that the dataset size equal to 500 images is the most efficient. Meanwhile, the experiment with the mosaic dataset shows constant accuracy improvement. Therefore, it is more advisable to collect different classes with 500 images than one large dataset. This study will be interesting not only for smart farming experts as well as for all machine learning experts.
Downloads
References
E. S. Ramil Brick, J. Holland, D. E. Anagnostou, K. Brown, and M. P. Y. Desmulliez, “A review of agroforestry, precision agriculture, and precision livestock farming—The case for a data-driven agroforestry strategy,” Frontiers in Sensors, vol. 3, Sep. 2022, doi: https://doi.org/10.3389/fsens.2022.998928.
Ultralytics, YOLOv8 GitHub repository. [Online] Available: https://github.com/ultralytics/ultralytics [Accessed: Jan15, 2024]
R. Bai, F. Shen, M. Wang, J. Lu, and Z. Zhang, “Improving Detection Capabilities of YOLOv8-n for Small Objects in Remote Sensing Imagery: Towards Better Precision with Simplified Model Complexity,” Jun. 2023, doi: https://doi.org/10.21203/rs.3.rs-3085871/v1.
Y. Bai, J. Yu, S. Yang, and J. Ning, “An improved YOLO algorithm for detecting flowers and fruits on strawberry seedlings,” Biosystems Engineering, vol. 237, pp. 1–12, Jan. 2024, doi: https://doi.org/10.1016/j.biosystemseng.2023.11.008.
S. Khalid, H. M. Oqaibi, M. Aqib, and Y. Hafeez, “Small Pests Detection in Field Crops Using Deep Learning Object Detection,” Sustainability, vol. 15, no. 8, p. 6815, Jan. 2023, doi: https://doi.org/10.3390/su15086815.
A. Elzebroek and K. Wind, Eds., “Guide to Cultivated Plants,” Jul. 2008, doi: https://doi.org/10.1079/9781845933562.0000.
I. Apeinans, L. Litavniece, S. Kodors, I. Zarembo, G. Lacis, and J. Deksne, “Smart Fruit Growing Through Digital Twin Paradigm: Systematic Review and Technology Gap Analysis,” Engineering Management in Production and Services, vol. 15, no. 4, pp. 128–143, Dec. 2023, doi: https://doi.org/10.2478/emj-2023-0033.
M. Arsenovic, M. Karanovic, S. Sladojevic, A. Anderla, and D. Stefanovic, “Solving Current Limitations of Deep Learning Based Approaches for Plant Disease Detection,” Symmetry, vol. 11, no. 7, p. 939, Jul. 2019, doi: https://doi.org/10.3390/sym11070939.
A. Bochkovskiy, C.-Y. Wang, and H.-Y. Liao, “YOLOv4: Optimal Speed and Accuracy of Object Detection,” Apr. 2020. Available: https://arxiv.org/pdf/2004.10934.pdf
Nicolaihaeni. (n.d.). GitHub - nicolaihaeni/MinneApple: A Benchmark Dataset for Apple Detection and Segmentation. [Online]. Available: https://github.com/nicolaihaeni/MinneApple [Accessed Jan. 3, 2024].
S. Kodors, M. Sondors, I. Apeinans, I. Zarembo, G. Lacis, E. Rubauskis and K. Karklina, “Importance of mosaic augmentation for agricultural image dataset”, 2024.
Downloads
Published
Issue
Section
License
Copyright (c) 2024 Ilmars Apeināns
This work is licensed under a Creative Commons Attribution 4.0 International License.
