Reduced dependence on chemical pesticides

Oct 13, 2025 by windailys.com 65 views

Introduction

Mango plants (Mangifera indica) are one of the most popular and economically significant fruits in Indonesia. However, the cultivation process of mango plants is often plagued by diseases and pests that can damage the leaves, reduce fruit production, and even threaten the survival of the plants. In an effort to overcome this problem, many farmers rely on pesticides, which can have devastating effects on the environment and human health. Therefore, a system that can utilize the color image and texture of the leaves to detect diseases and pests in mango leaves quickly and accurately is urgently needed.

Background

The use of pesticides in agriculture has been a long-standing issue, as it can contaminate soil, water, and air, and even affect human health. Moreover, the overuse of pesticides can lead to the development of pesticide-resistant pests, making it even more challenging to control them. In recent years, there has been a growing interest in developing alternative methods for pest control, such as using integrated pest management (IPM) techniques, which involve a combination of physical, cultural, biological, and chemical controls.

Methodology

In this study, the You Only Look Once method version 5 (Yolov5) was used to detect the types of diseases and pests in mango leaves in real-time. The dataset used in this study consisted of 1,250 data points, with 1,000 data points for training and 250 data points for validation. For testing, 120 data points were taken using a smartphone camera. The implementation of the Yolov5 method in detecting disease and pests in mango leaves managed to reach an accuracy level of 93.3%.

Implementation of Yolov5

The Yolov5 method is a deep learning-based approach that uses a convolutional neural network (CNN) to detect objects in images. The method consists of three main components: feature extraction, object detection, and post-processing. The feature extraction component uses a series of convolutional and pooling layers to extract features from the input image. The object detection component uses a region proposal network (RPN) to generate region proposals, which are then classified as either a pest or a non-pest. The post-processing component refines the detection results by applying non-maximum suppression and class activation mapping.

Results and Discussion

The results of this study show that the Yolov5 method is effective in detecting the types of diseases and pests in mango leaves. The accuracy level of 93.3% indicates that the system developed by the Yolov5 method is reliable and can be used in real-world applications. The use of Yolov5 in pest detection and disease in mango leaves provides significant benefits compared to traditional methods. First, this system offers speed in detecting problems, which allows farmers to take action immediately. In the context of agriculture, time is key; early detection can reduce damage caused by pests and diseases.

Second, this approach reduces dependence on chemical pesticides. By detecting problems early, farmers can apply more precise controls, such as using vegetable pesticides or more environmentally friendly organic agricultural techniques. This is very important in the context of agricultural sustainability, where soil and ecosystem health must be maintained.

Conclusion

The use of Yolov5 in pest detection and disease in mango leaves shows great potential to be implemented on a wider scale. With further development and increasing dataset, this system can be a reliable tool for farmers in maintaining their mango plants. Finally, the use of technology such as Yolov5 not only increases efficiency and effectiveness in agricultural management, but also contributes to efforts to protect the environment and public health. With innovative steps like this, it is hoped that the Indonesian agricultural industry can develop sustainably, while meeting the needs of the market that continues to increase.

Future Work

There are several areas that need to be explored in future work. First, the dataset used in this study is relatively small, and more data points are needed to improve the accuracy of the system. Second, the Yolov5 method can be improved by using more advanced techniques, such as transfer learning and ensemble learning. Third, the system can be integrated with other technologies, such as drones and satellite imaging, to provide more accurate and comprehensive information on pest and disease management.

References

[1] You Only Look Once: Unified, Real-Time Object Detection by J. Redmon, A. Farhadi, Y. Liu, and D. Anguelov. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016.
[2] Deep Learning for Computer Vision: A Review by Y. LeCun, Y. Bengio, and G. Hinton. In Proceedings of the IEEE, 2015.
[3] Pest and Disease Management in Mango Plants: A Review by S. K. Singh, R. K. Singh, and A. K. Singh. In Journal of Plant Protection, 2018.

Appendix

Q: What is the You Only Look Once method version 5 (Yolov5)?

A: The You Only Look Once method version 5 (Yolov5) is a deep learning-based approach that uses a convolutional neural network (CNN) to detect objects in images. It is a real-time object detection system that can be used for various applications, including pest and disease detection in mango plants.

Q: How does the Yolov5 method work?

A: The Yolov5 method works by using a series of convolutional and pooling layers to extract features from the input image. The extracted features are then used to generate region proposals, which are classified as either a pest or a non-pest. The final detection results are obtained by applying non-maximum suppression and class activation mapping.

Q: What are the benefits of using the Yolov5 method for pest and disease detection in mango plants?

A: The benefits of using the Yolov5 method for pest and disease detection in mango plants include:

Speed: The Yolov5 method can detect pests and diseases in real-time, allowing farmers to take action immediately.
Accuracy: The Yolov5 method has an accuracy level of 93.3%, making it a reliable tool for farmers.
Reduced dependence on chemical pesticides: The Yolov5 method can help farmers reduce their dependence on chemical pesticides by detecting problems early and allowing them to apply more precise controls.

Q: What are the limitations of the Yolov5 method?

A: The limitations of the Yolov5 method include:

Small dataset: The dataset used in this study is relatively small, and more data points are needed to improve the accuracy of the system.
Limited generalizability: The Yolov5 method may not be generalizable to other types of plants or pests.

Q: How can the Yolov5 method be improved?

A: The Yolov5 method can be improved by:

Increasing the dataset size: More data points are needed to improve the accuracy of the system.
Using more advanced techniques: Techniques such as transfer learning and ensemble learning can be used to improve the accuracy of the system.
Integrating with other technologies: The Yolov5 method can be integrated with other technologies, such as drones and satellite imaging, to provide more accurate and comprehensive information on pest and disease management.

Q: What are the future applications of the Yolov5 method?

A: The future applications of the Yolov5 method include:

Pest and disease detection in other types of plants: The Yolov5 method can be used for pest and disease detection in other types of plants, such as coffee, cocoa, and tea.
Precision agriculture: The Yolov5 method can be used for precision agriculture, allowing farmers to apply more precise controls and reduce waste.
Environmental monitoring: The Yolov5 method can be used for environmental monitoring, allowing researchers to track changes in the environment and detect early warning signs of environmental degradation.

Q: How can readers get involved in the development of the Yolov5 method?

A: Readers can get involved in the development of the Yolov5 method by:

Contributing to the dataset: More data points are needed to improve the accuracy of the system. Readers can contribute to the dataset by collecting and labeling images of pests and diseases in mango plants.
Sharing their expertise: Readers with expertise in deep learning, computer vision, or agriculture can share their expertise and help improve the accuracy of the system.
Participating in the development process: Readers can participate in the development process by contributing to the code, testing the system, and providing feedback.