Strawberry Disease Detection Using Multispectral UAV Imagery
Accurate, timely, and non-invasive diagnosis of plant diseases is essential in the industrial cultivation of strawberries, as it helps minimize yield losses and reduce treatment costs. With the advancement of unmanned aerial vehicles and sensor technologies, remote sensing has emerged as a promising...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | Russian |
| Published: |
Federal Scientific Agroengineering Centre VIM
2025-07-01
|
| Series: | Сельскохозяйственные машины и технологии |
| Subjects: | |
| Online Access: | https://www.vimsmit.com/jour/article/view/667 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Accurate, timely, and non-invasive diagnosis of plant diseases is essential in the industrial cultivation of strawberries, as it helps minimize yield losses and reduce treatment costs. With the advancement of unmanned aerial vehicles and sensor technologies, remote sensing has emerged as a promising tool for monitoring crop health and detecting diseases. Early detection is especially important for sensitive crops such as garden strawberries. (Research purpose) The research aims to evaluate the potential of using multispectral sensors and unmanned aerial vehicles for detecting fungal diseases in garden strawberries under field conditions. (Materials and methods) Aerial imaging of strawberry plants affected by white leaf spot was carried out at the experimental nursery of the Siberian Federal Research Center for Agro-BioTechnologies of the RAS (SibFRC ABT RAS). A DJI Phantom 4 Multispectral quadcopter equipped with a multispectral camera was used for data collection. The acquired imagery underwent preliminary processing, including orthophotomap generation and extraction of spectral and textural features from the images. (Results and discussion) Analysis of the multispectral data enabled the identification of informative feature sets for distinguishing between healthy and fungus-infected strawberry plants. A disease detection model developed using the Random Forest algorithm, achieved a classification accuracy of 77%. (Conclusions) To improve classification accuracy, further research involving sensors with higher spatial resolution is recommended. Another promising direction is the development of classification models based on convolutional neural networks, which offer improved performance through deeper image analysis. The results confirm the potential of UAV-based multispectral imaging for effective crop disease monitoring. |
|---|---|
| ISSN: | 2073-7599 |