A Dataset of Visible Light and Thermal Infrared Images for Health Monitoring of Caged Laying Hens in Large-scale Farming DOI Open Access
Weihong Ma, Xingmeng Wang,

Xianglong Xue

et al.

Published: Aug. 21, 2024

Considering animal welfare, the free-range laying hen farming model is increasingly gaining attention. However, in some countries, large-scale still relies on cage-rearing model, making focus welfare of caged hens equally important. To evaluate health status hens, a dataset comprising visible light and thermal infrared images was established for analyses, including morphological, thermographic, comb, behavioural as-sessments, enabling comprehensive evaluation hens' health, behaviour, population counts. address issue insufficient data samples detection process indi-vidual group named BClayinghens constructed containing 61,133 images. The completed using three types devices: smartphones, cameras, cameras. All correspond to have achieved positional alignment through coordinate correction. Additionally, were annotated with chicken head labels, obtaining 63,693 which can be directly used training deep learning models object combined corresponding analyze temperature heads. enable deep-learning recognition adapt different breeding environ-ments, various enhancement methods such as rotation, shearing, colour enhancement, noise addition image processing. important ap-plying detection, analysis, counting under farming.

Language: Английский

Only Detect Broilers Once (ODBO): A Method for Monitoring and Tracking Individual Behavior of Cage-Free Broilers DOI Creative Commons

Chengcheng Yin,

Xinjie Tan,

Xiaoxin Li

et al.

Agriculture, Journal Year: 2025, Volume and Issue: 15(7), P. 669 - 669

Published: March 21, 2025

In commercial poultry farming, automated behavioral monitoring systems hold significant potential for optimizing production efficiency and improving welfare outcomes at scale. The detection of free-range broilers matters precision farming animal welfare. Current research often focuses on either behavior or individual tracking, with few studies exploring their connection. To continuously track broiler behaviors, the Only Detect Broilers Once (ODBO) method is proposed by linking behaviors identity information. This has a detector, an Tracker, Connector. First, integrating SimAM, WIOU, DIOU-NMS into YOLOv8m, high-performance YOLOv8-BeCS detector created. It boosts P 6.3% AP 3.4% compared to original detector. Second, designed Connector, based tracking-by-detection structure, transforms tracking task, combining recognition. Tests sort-series trackers show HOTA, MOTA, IDF1 increase 27.66%, 28%, 27.96%, respectively, after adding Fine-tuning experiments verify model’s generalization. results this outperforms others in accuracy, generalization, convergence speed, providing effective behaviors. addition, system’s ability simultaneously monitor bird indicators group dynamics could enable data-driven decisions management.

Language: Английский

Citations

1
Equivalence Between Optical Flow, the Unrest Index, and Walking Distance to Estimate the Welfare of Broiler Chickens DOI Creative Commons
Danilo Florentino Pereira, Irenilza de Alencar Nääs, Saman Abdanan Mehdizadeh

et al.

Animals, Journal Year: 2025, Volume and Issue: 15(9), P. 1311 - 1311

Published: May 1, 2025

Modern poultry production demands scalable and non-invasive methods to monitor animal welfare, particularly as broiler strains are increasingly bred for rapid growth, often at the expense of mobility health. This study evaluates two advanced computer vision techniques—Optical Flow Unrest Index—to assess movement patterns in chickens. Three commercial (Hybro®, Cobb®, Ross®) were housed controlled environments continuously monitored using ceiling-mounted video systems. Chicken movements detected tracked a YOLO model, with centroid data informing both Index distance walked metrics. Optical velocity metrics (mean, variance, skewness, kurtosis) extracted Farnebäck algorithm. Pearson correlation analyses revealed strong associations between variables traditional indicators, average showing strongest Index. Among evaluated strains, Cobb® demonstrated variance Index, indicating distinct profile. The equipment’s camera’s slight instability had minimal effect on measurement. Still, its walking accredits it an effective method high-resolution behavioral monitoring. supports integration technologies into precision livestock systems, offering foundation predictive welfare management scale.

Language: Английский

Citations

0
Edge AI-enabled chicken health detection based on enhanced FCOS-Lite and knowledge distillation DOI
Q.K. Tong, Jinrui Wang, Wenshuang Yang

et al.

Computers and Electronics in Agriculture, Journal Year: 2024, Volume and Issue: 226, P. 109432 - 109432

Published: Sept. 10, 2024

Language: Английский

Citations

3
Automatic monitoring of activity intensity in a chicken flock using a computer vision-based background image subtraction technique: an experimental infection study with fowl adenovirus DOI Creative Commons
Hiroshi Iseki, Eri Furukawa, Toshiaki Shimasaki

et al.

Smart Agricultural Technology, Journal Year: 2025, Volume and Issue: unknown, P. 100821 - 100821

Published: Feb. 1, 2025

Language: Английский

Citations

0
Review: multi object tracking in livestock - from farm animal management to state-of-the-art methods DOI Creative Commons

Malik Nidhi,

Kai Liu,

K J Flay

et al.

animal, Journal Year: 2025, Volume and Issue: unknown, P. 101503 - 101503

Published: April 1, 2025

Language: Английский

Citations

0
Optimizing Deep Learning Algorithms for Effective Chicken Tracking through Image Processing DOI Creative Commons
Saman Abdanan Mehdizadeh, Allan Lincoln Rodrigues Siriani, Danilo Florentino Pereira

et al.

AgriEngineering, Journal Year: 2024, Volume and Issue: 6(3), P. 2749 - 2767

Published: Aug. 8, 2024

Identifying bird numbers in hostile environments, such as poultry facilities, presents significant challenges. The complexity of these environments demands robust and adaptive algorithmic approaches for the accurate detection tracking birds over time, ensuring reliable data analysis. This study aims to enhance methodologies automated chicken identification videos, addressing dynamic non-standardized nature farming environments. YOLOv8n model was chosen due its high portability. developed algorithm promptly identifies labels chickens they appear image. process is illustrated two parallel flowcharts, emphasizing different aspects image processing behavioral False regions chickens’ heads tails are excluded calculate body area more accurately. following three scenarios were tested with newly modified deep-learning algorithm: (1) reappearing temporary invisibility; (2) multiple missing object occlusion; (3) coalescing chickens. results a precise measure size shape, YOLO achieving an accuracy above 0.98 loss less than 0.1. In all scenarios, improved maintaining identification, enabling simultaneous several respective error rates 0, 0.007, 0.017. Morphological based on features extracted from each chicken, proved be effective strategy enhancing accuracy.

Language: Английский

Citations

1
A Dataset of Visible Light and Thermal Infrared Images for Health Monitoring of Caged Laying Hens in Large-Scale Farming DOI Creative Commons
Weihong Ma, Xingmeng Wang,

Xianglong Xue

et al.

Sensors, Journal Year: 2024, Volume and Issue: 24(19), P. 6385 - 6385

Published: Oct. 2, 2024

Considering animal welfare, the free-range laying hen farming model is increasingly gaining attention. However, in some countries, large-scale still relies on cage-rearing model, making focus welfare of caged hens equally important. To evaluate health status hens, a dataset comprising visible light and thermal infrared images was established for analyses, including morphological, thermographic, comb, behavioral assessments, enabling comprehensive evaluation hens’ health, behavior, population counts. address issue insufficient data samples detection process individual group named BClayinghens constructed containing 61,133 images. The completed using three types devices: smartphones, cameras, cameras. All correspond to have achieved positional alignment through coordinate correction. Additionally, were annotated with chicken head labels, obtaining 63,693 which can be directly used training deep learning models object combined corresponding analyze temperature heads. enable deep-learning recognition adapt different breeding environments, various enhancement methods such as rotation, shearing, color enhancement, noise addition image processing. important applying detection, analysis, counting under farming.

Language: Английский

Citations

1
Autonomous inspection robot for dead laying hens in caged layer house DOI
Weihong Ma, Xingmeng Wang, Simon X. Yang

et al.

Computers and Electronics in Agriculture, Journal Year: 2024, Volume and Issue: 227, P. 109595 - 109595

Published: Nov. 9, 2024

Language: Английский

Citations

1
A Dataset of Visible Light and Thermal Infrared Images for Health Monitoring of Caged Laying Hens in Large-scale Farming DOI Open Access
Weihong Ma, Xingmeng Wang,

Xianglong Xue

et al.

Published: Aug. 21, 2024

Considering animal welfare, the free-range laying hen farming model is increasingly gaining attention. However, in some countries, large-scale still relies on cage-rearing model, making focus welfare of caged hens equally important. To evaluate health status hens, a dataset comprising visible light and thermal infrared images was established for analyses, including morphological, thermographic, comb, behavioural as-sessments, enabling comprehensive evaluation hens' health, behaviour, population counts. address issue insufficient data samples detection process indi-vidual group named BClayinghens constructed containing 61,133 images. The completed using three types devices: smartphones, cameras, cameras. All correspond to have achieved positional alignment through coordinate correction. Additionally, were annotated with chicken head labels, obtaining 63,693 which can be directly used training deep learning models object combined corresponding analyze temperature heads. enable deep-learning recognition adapt different breeding environ-ments, various enhancement methods such as rotation, shearing, colour enhancement, noise addition image processing. important ap-plying detection, analysis, counting under farming.

Language: Английский

Citations

1