BRA-YOLOv7: Improved Detection of Tea Leaf Diseases
Author Information
Author(s): Ye Rong, Gao Quan, Li Tong
Primary Institution: Yunnan Agricultural University
Hypothesis
Can the BRA-YOLOv7 model improve the detection accuracy of tea leaf diseases in complex environments?
Conclusion
The BRA-YOLOv7 model significantly enhances the detection accuracy and efficiency of tea leaf diseases compared to traditional models.
Supporting Evidence
- The BRA-YOLOv7 model achieved a 4.8% improvement in recognition accuracy over YOLOv7.
- Improvements in recall rate and balance score were 5.3% and 5%, respectively.
- The model reduced floating-point operations by 1.4G and improved FPS by 5.52%.
Takeaway
This study created a new model to help find diseases on tea leaves better and faster, which is important for growing healthy tea.
Methodology
The study used a dataset of tea disease images and compared the performance of the BRA-YOLOv7 model against other models like YOLOv6 and YOLOv7 using metrics such as mAP and FPS.
Potential Biases
Potential bias may arise from the reliance on specific lighting conditions and the manual annotation process.
Limitations
The study may be limited by the dataset's diversity and the environmental conditions under which the images were captured.
Participant Demographics
The study focused on tea plantations in Yunnan, China.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website