GSD-YOLO: A Lightweight Decoupled Wheat Scab Spore Detection Network Based on Yolov7-Tiny DOI Creative Commons
Dongyan Zhang,

Wenfeng Tao,

Tao Cheng

et al.

Agriculture, Journal Year: 2024, Volume and Issue: 14(12), P. 2278 - 2278

Published: Dec. 12, 2024

Aimed at the problem of difference between intra-class and inter-class pathogenic spores Wheat Scab image being small difficult to distinguish, in this paper, we propose a lightweight decoupled spore detection network based on Yolov7-tiny (GSD-YOLO). Specifically, considering limitations storage space power consumption actual field equipment, original head is optimized as head, GSConv module embedded reduce parameters model number calculations required. In addition, utilize an improved Spore–Copy data augmentation strategy improve performance generalization ability algorithm fit large numbers, morphology, variety wheat disease efficiency constructing set diverse spores. The experimental results show that mAP proposed reaches 98.0%, which 3.9 percentage points higher than model. At same time, speed 114 f/s, memory 13.1 MB, meets application requirements hardware deployment real-time detection. It can provide some technical support prevention grading farmland.

Language: Английский

DAD-YOLO as a novel computer vision tool to predict the environmental impact of harmful algae presence in contaminated river water employed for large-scale irrigation to agricultural field DOI

S.S. Jayakrishna,

S. Sankar Ganesh

Journal of Water Process Engineering, Journal Year: 2025, Volume and Issue: 71, P. 107439 - 107439

Published: March 1, 2025

Language: Английский

Citations

2
High-throughput phenotyping techniques for forage: Status, bottleneck, and challenges DOI Creative Commons
Tao Cheng, Dongyan Zhang, Gan Zhang

et al.

Artificial Intelligence in Agriculture, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 1, 2025

Language: Английский

Citations

0
ECL-Tear: Lightweight detection method for multiple types of belt tears DOI
Xiaopan Wang, Shuting Wan,

Zhonghang Li

et al.

Measurement, Journal Year: 2025, Volume and Issue: unknown, P. 117269 - 117269

Published: March 1, 2025

Language: Английский

Citations

0
A LIGHTWEIGHT MILLET DOWNY MILDEW SPORE DETECTION METHOD BASED ON IMPROVED YOLOv8s DOI
Jun Yan, Zhaohui Zhai, Zhiyuan Feng

et al.

INMATEH Agricultural Engineering, Journal Year: 2025, Volume and Issue: unknown, P. 323 - 333

Published: April 9, 2025

This paper proposes a lightweight spore detection method for millet downy mildew based on an improved YOLOv8s, aiming to enhance the accuracy and efficiency of detection. First, backbone network YOLOv8s model was modified by replacing original with EfficientViT. The substitution EfficientViT enables global receptive field multi-scale learning, which helps reduce computational costs. While maintaining high performance, this modification significantly improves efficiency. Second, Frequency-Adaptive Dilated Convolution (FADC) module added neck model. By adaptively adjusting dilated convolution, FADC optimizes different frequency information. It small objects without adding extra burden. Finally, head optimized better adapt task detecting spores, resulting in enhanced speed accuracy. algorithm, named EFP-YOLOv8s, maintains same mAP50 as while reducing number parameters 37.8% cost 58.5%. balancing performance reduced resource demands, achieves design, making it more deployable scalable practical applications.

Language: Английский

Citations

0
GSD-YOLO: A Lightweight Decoupled Wheat Scab Spore Detection Network Based on Yolov7-Tiny DOI Creative Commons
Dongyan Zhang,

Wenfeng Tao,

Tao Cheng

et al.

Agriculture, Journal Year: 2024, Volume and Issue: 14(12), P. 2278 - 2278

Published: Dec. 12, 2024

Aimed at the problem of difference between intra-class and inter-class pathogenic spores Wheat Scab image being small difficult to distinguish, in this paper, we propose a lightweight decoupled spore detection network based on Yolov7-tiny (GSD-YOLO). Specifically, considering limitations storage space power consumption actual field equipment, original head is optimized as head, GSConv module embedded reduce parameters model number calculations required. In addition, utilize an improved Spore–Copy data augmentation strategy improve performance generalization ability algorithm fit large numbers, morphology, variety wheat disease efficiency constructing set diverse spores. The experimental results show that mAP proposed reaches 98.0%, which 3.9 percentage points higher than model. At same time, speed 114 f/s, memory 13.1 MB, meets application requirements hardware deployment real-time detection. It can provide some technical support prevention grading farmland.

Language: Английский

Citations

0