The impact of plant and fruit diseases on agricultural economies is significant, resulting in the reduction of crop quality and yield. The development of precise and automated detection techniques to overcome this issue is crucial in order to minimize agricultural losses and foster economic growth. A deep learning approach named YOLO-AppleScab is introduced, Which integrates the Content-Aware ReAssembly of Feature ( ) architecture into YOLOv7 to enhance the detection of apple fruits with healthy and disease classification. Moreover, the model used the traditional bounding box (R-Bbox) for apple fruit localization. The performance metrics of the proposed model were noteworthy, with F1, recall, and precision rates of 89.75%, 85.20%, and 94.80%, respectively. Furthermore, the results demonstrate a mean average precision ( ) of 89.30% when evaluated at an intersection over a union (IoU) threshold of 0.5. Additionally, the model achieved a of 64% with an average duration of inference per image of 175.2 milliseconds. The integration of the YOLOv7 head was a crucial factor in attaining superior detection capabilities in comparison to contemporary techniques. The research highlights the importance of utilizing deep-learning methodologies for accurate and automated detection of apple scab disease, which has potential benefits for reducing agricultural damages and promoting economic development.
Previous Article in event
Previous Article in session
Next Article in event
Next Article in session
YOLO-AppleScab: A Deep Learning Approach for Efficient and Accurate Apple Scab Detection in Varied Lighting Conditions Using CARAFE-enhanced YOLOv7
Published:
29 December 2023
by MDPI
in 2nd International Online Conference on Agriculture
session Artificial Intelligence for Advanced Analyses in Agriculture;
Abstract:
Keywords: Scab disease; CARAFE; YOLO-AppleScab; bounding box; means average precision; average inference time per image; disease detection.
![Top](/images/top.png?version=d2a1072bcb778c072486fad356669189)