YOLOv7-DWS: tea bud recognition and detection network in multi-density environment via improved YOLOv7

Frontiers in Plant Science, Journal Year: 2025, Volume and Issue: 15

Published: Jan. 7, 2025

Accurate detection and recognition of tea bud images can drive advances in intelligent harvesting machinery for gardens technology pests diseases. In order to realize the grading buds a complex multi-density garden environment. This paper proposes an improved YOLOv7 object algorithm, called YOLOv7-DWS, which focuses on improving accuracy recognition. First, we make series improvements including decouple head replace YOLOv7, enhance feature extraction ability model optimize class decision logic. The problem simultaneous classification one-bud-one-leaf one-bud-two-leaves was solved. Secondly, new loss function WiseIoU is proposed improves model. Finally, evaluate different attention mechanisms model's focus key features. experimental results show that algorithm has significantly over original all evaluation indexes, especially RTea (+6.2%) [email protected] (+7.7%). From results, this helps provide perspective possibility field image

Language: Английский

---

Enhanced Methodology and Experimental Research for Caged Chicken Counting Based on YOLOv8

Animals, Journal Year: 2025, Volume and Issue: 15(6), P. 853 - 853

Published: March 16, 2025

Accurately counting chickens in densely packed cages is a major challenge large-scale poultry farms. Traditional manual methods are labor-intensive, costly, and prone to errors due worker fatigue. Furthermore, current deep learning models often struggle with accuracy caged environments because they not well-equipped handle occlusions. In response, we propose the You Only Look Once-Chicken Counting Algorithm (YOLO-CCA). YOLO-CCA improves YOLOv8-small model by integrating CoordAttention mechanism Reversible Column Networks backbone. This enhancement improved model’s F1 score 96.7% (+3%) average precision50:95 80.6% (+2.8%). Additionally, developed threshold-based continuous frame inspection method that records maximum number of per cage corresponding timestamps. The data stored cloud database for reliable tracking during robotic inspections. experiments were conducted an actual farming environment, involving 80 total 493 chickens, showed raised chicken recognition rate 90.9% (+13.2%). When deployed on Jetson AGX Orin industrial computer using TensorRT, detection speed increased 90.9 FPS (+57.6 FPS), although slightly decreased 93.2% (−2.9%). summary, reduces labor costs, efficiency, supports intelligent transformation.

Language: Английский

---

A Lightweight Model for Small-Target Pig Eye Detection in Automated Estrus Recognition

Animals, Journal Year: 2025, Volume and Issue: 15(8), P. 1127 - 1127

Published: April 13, 2025

In modern large-scale pig farming, accurately identifying sow estrus and ensuring timely breeding are crucial for maximizing economic benefits. However, the short duration of reliance on subjective human judgment pose significant challenges precise insemination timing. To enable non-contact, automated detection, this study proposes an improved algorithm, Enhanced Context-Attention YOLO (ECA-YOLO), based YOLOv11. The model utilizes ocular appearance features—eye’s spirit, color, shape, morphology—across different stages as key indicators. MSCA module enhances small-object detection efficiency, while PPA GAM modules improve feature extraction capabilities. Additionally, Adaptive Threshold Focal Loss (ATFL) function increases model’s sensitivity to hard-to-classify samples, enabling accurate stage classification. was trained validated a dataset comprising 4461 images eyes during benchmarked against YOLOv5n, YOLOv7tiny, YOLOv8n, YOLOv10n, YOLOv11n, Faster R-CNN. Experimental results demonstrate that ECA-YOLO achieves mean average precision (mAP) 93.2%, F1-score 88.0%, with 5.31M parameters, FPS reaches 75.53 frames per second, exhibiting superior overall performance. findings confirm feasibility using features highlight potential real-time, monitoring under complex farming conditions. This lays groundwork in intensive farming.

Language: Английский

---

Heat Stress Influences Immunity Through DUSP1 and HSPA5 Mediated Antigen Presentation in Chickens

Animals, Journal Year: 2025, Volume and Issue: 15(8), P. 1141 - 1141

Published: April 16, 2025

The objective of this study was to elucidate the immune system response heat stress in chickens. In study, mRNA-seq conducted on spleen and bursa experimental chickens, six differentially expressed genes associated with immunity were present following immunization. Following exposure stress, 15 related shock proteins identified. Furthermore, expression levels DUSP1 HSPA5 significantly lower non-stressed group. With regard mechanism, overexpression or resulted no significant difference MHC-I, MHC-II, CD80 mRNA expression. However, stimulation LPS, CD80, CD86, CD1C, IL1B, TLR4 increased. enhancement observed occur at an earlier stage than when LPS stimulated alone, thereby facilitating recognition by HD11. inhibition alterations detected. CD1C notably diminished. conclusion, have been demonstrated play important roles affecting antigen presentation. provides a theoretical basis for regulation mechanism disease resistance poultry.

Language: Английский

---