An improved small object detection CTB-YOLO model for early detection of tip-burn and powdery mildew symptoms in coriander (Coriandrum sativum) for indoor environment using an edge device
Indoor cultivation of coriander (Coriandrum sativum) ensures a steady supply to meet year-round consumer demand; however, tip-burn and powdery mildew significantly challenge indoor coriander production. Early and accurate detection of these symptoms is critical for maintaining yield and quality, yet...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-12-01
|
| Series: | Smart Agricultural Technology |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2772375525003740 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Indoor cultivation of coriander (Coriandrum sativum) ensures a steady supply to meet year-round consumer demand; however, tip-burn and powdery mildew significantly challenge indoor coriander production. Early and accurate detection of these symptoms is critical for maintaining yield and quality, yet traditional visual inspection methods are subjective and prone to error. To address this, we developed an enhanced deep learning model, Coriander Tip-Burn YOLO (CTB-YOLO), specifically tailored for detecting small-object symptoms in coriander leaves, emphasizing the reduction of false-positive detections (FP), which could lead to erroneous alerts and unnecessary interventions. We created a novel, annotated dataset comprising 3240 images of tip-burn and 3340 images of powdery mildew symptoms collected under controlled indoor conditions. The CTB-YOLO model incorporates advanced multiscale feature fusion, significantly improving detection accuracy with mean average precision (mAP) scores of 76.1% for tip-burn and 69.3% for powdery mildew, while notably reducing false-positive detections. The model was deployed on an edge computing device integrated with the LINE application, providing real-time notifications to growers for immediate and reliable intervention. This research demonstrates CTB-YOLO’s potential as an effective tool for automating disease monitoring in indoor agriculture, enhancing crop health management, and promoting sustainable farming practices. |
|---|---|
| ISSN: | 2772-3755 |