Review: Smart agri-systems for the pig industry

How do pig veterinarians view technology-assisted data utilisation for pig health and welfare management? A qualitative study in Spain, the Netherlands, and Ireland

BACKGROUND

Application of data-driven strategies may support veterinarians' decision-making, benefitting pig disease prevention and control. However, little is known about veterinarians' need for data utilisation to support their decision-making process. The current study used qualitative methods, specifically focus group discussions, to explore veterinarians' views on data utilisation and their need for data tools in relation to pig health and welfare management in Spain, the Netherlands, and Ireland.

RESULTS

Generally, veterinarians pointed out the potential benefits of using technology for pig health and welfare management, but data is not yet structurally available to support their decision-making. Veterinarians pointed out the challenge of collecting, recording, and accessing data in a consistent and timely manner. Besides, the reliability, standardisation, and the context of data were identified as important factors affecting the efficiency and effectiveness of data utilisation by veterinarians. A user-friendly, adaptable, and integrated data tool was regarded as potentially helpful for veterinarians' daily work and supporting their decision-making. Specifically, veterinarians, particularly independent veterinary practitioners, noted a need for easy access to pig information. Veterinarians such as those working for integrated companies, corporate veterinarians, and independent veterinary practitioners expressed their need for data tools that provide useful information to monitor pig health and welfare in real-time, to visualise the prevalence of endemic disease based on a shared report between farmers, veterinarians, and other professional parties, to support decision-making, and to receive early warnings for disease prevention and control.

CONCLUSIONS

It is concluded that the management of pig health and welfare may benefit from data utilisation if the quality of data can be assured, the data tools can meet veterinarians' needs for decision-making, and the collaboration of sharing data and using data between farmers, veterinarians, and other professional parties can be enhanced. Nevertheless, several notable technical and institutional barriers still exist, which need to be overcome.

Identification and functional prediction of miRNAs that regulate ROS levels in dielectric barrier discharge plasma-treated boar spermatozoa

Dielectric barrier discharge (DBD) plasma regulates the levels of reactive oxygen species (ROS), which are critical for sperm quality. MicroRNAs (miRNAs) are non-coding single-stranded RNA molecules encoded by endogenous genes, which regulate post-transcriptional gene expression in animals. At present, it is unknown whether DBD plasma can regulate sperm ROS levels through miRNAs. To further understand the regulatory mechanism of DBD plasma on sperm ROS levels, miRNAs in fresh boar spermatozoa were detected using Illumina deep sequencing technology. We found that 25 known miRNAs and 50 novel miRNAs were significantly upregulated, and 14 known miRNAs and 74 novel miRNAs were significantly downregulated in DBD plasma-treated spermatozoa. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis showed that target genes of differentially expressed miRNAs were involved in many activities and pathways associated with antioxidants. We verified that DBD plasma significantly increased boar sperm quality and reduced ROS levels. These results suggest that DBD plasma can improve sperm quality by regulating ROS levels via miRNAs. Our findings provide a potential strategy to improve sperm quality through miRNA-targeted regulation of ROS, which helps to increase male reproduction and protect cryopreserved semen in clinical practice.

GooseDetectlion: A Fully Annotated Dataset for Lion-head Goose Detection in Smart Farms

Large datasets are required to develop Artificial Intelligence (AI) models in AI powered smart farming for reducing farmers' routine workload, this paper contributes the first large lion-head goose dataset GooseDetectlion, which consists of 2,660 images and 98,111 bounding box annotations. The dataset was collected with 6 cameras deployed in a goose farm in Chenghai district of Shantou city, Guangdong province, China. Images sampled from videos collected during July 9 -10 in 2022 were fully annotated by a team of fifty volunteers. Compared with another 6 well known animal datasets in literature, our dataset has higher capacity and density, which provides a challenging detection benchmark for main stream object detectors. Six state-of-the-art object detectors have been selected to be evaluated on the GooseDetectlion, which includes one two-stage anchor-based detector, three one-stage anchor-based detectors, as well as two one-stage anchor-free detectors. The results suggest that the one-stage anchor-based detector You Only Look Once version 5 (YOLO v5) achieves the best overall performance in terms of detection precision, model size and inference efficiency.

Effects of parity, season of birth, and sex on within-litter variation and pre-weaning performance of F1 Large White × Landrace pigs

Background and Aim

A piglet's pre-weaning performance significantly influences both animal welfare and profitability in pig production. Understanding piglet pre-weaning performance influencing factors is key to enhancing animal welfare, reducing losses, and boosting profitability. The study aimed to evaluate the impact of parity, season of birth, and sex on within-litter variation and pre-weaning performance of F1 Large White × Landrace pigs.

Materials and Methods

Information regarding total litter size, number of born alive, number of stillbirths, piglet weight at birth, mortality, and count of weaned F1 Large White × Landrace piglets was acquired from the farm database (April 2022-February 2023). 2602 females and 2882 males, a total of 5484 piglets were utilized, with records from 360 sows. The coefficient of variation (CV) of birth weights among piglets within a litter was calculated. The general linear model analysis in MiniTab 17 was used to evaluate the data, with Fisher's least significant difference test (p < 0.05) used for mean separation and Pearson's moment correlation coefficient calculated to assess relationships between survival rates, mortality rates, litter size, birth weight, and birth weight CV.

Results

Parity had a statistically significant impact on litter size, birth weight, and survival rate (p < 0.05). The sow's parity did not significantly (p > 0.05) impact the number of piglets born alive or weaned. Multiparous sows had a significantly larger litter size (p < 0.05) than primiparous sows at birth. The litter weights for parities 2, 4, and 5 did not significantly differ (p > 0.05), with averages of 20.95, 20.74, and 20.03 kg, respectively. About 91.29% was the highest survival rate recorded in parity 2 (p < 0.05). The 1st week of life recorded an 8.02% mortality rate. The mortality rate in parity 3-5 group was significantly (p < 0.05) higher (11.90%) in week 1 than in the other groups (parity 1: 6.79%, parity 2: 5.74%, parity 3-5: 8.54 and 9.21%). The litter sizes in autumn (17.34) and spring (17.72) were significantly larger (p < 0.05) than those in summer (16.47) and winter (16.83). In autumn and spring, the survival rate (83.15 and 85.84%, respectively) was significantly lower (p < 0.05) compared to summer (88.40%) and winter (89.07%). In all seasons, the litter weights did not significantly differ (p > 0.05). The birth weight CV was significantly (p < 0.05) lower during summer (20.11%) than during spring (22.43%), autumn (23.71%), and winter (21.69%). The season of birth had no significant effect (p > 0.05) on the number of live piglets. Males (1.34 kg) were heavier (p < 0.05) than females (1.30 kg) at birth. Notably, the birth weight CV was similar between males (22.43%) and females (22.52%). Litter size was positively correlated with average litter weight (rp = 0.576, p < 0.001), birth weight CV (rp = 0.244, p < 0.001), and mortality rate (rp = 0.378, p < 0.001). An insignificant relationship was observed between average litter weight and birth weight CV (rp = -0.028, p > 0.05) and survival rate (rp = -0.032, p > 0.05).

Conclusion

In F1 Large White × Landrace pigs, birth uniformity among piglets declines as litter size grows larger. In parity 3-5, multiparous sows yield litters with reduced uniformity. With an increase in litter size, uniformity among piglets at birth worsens. A larger litter size and greater piglet birth weight variation are linked to a higher pre-weaning mortality rate. Producers need a balanced selection approach to boost litter size and must cull aging sows carefully to introduce younger, more productive females.

Mulberry Leaf Dietary Supplementation Can Improve the Lipo-Nutritional Quality of Pork and Regulate Gut Microbiota in Pigs: A Comprehensive Multi-Omics Analysis

Mulberry leaves, a common traditional Chinese medicine, represent a potential nutritional strategy to improve the fat profile, also known as the lipo-nutrition, of pork. However, the effects of mulberry leaves on pork lipo-nutrition and the microorganisms and metabolites in the porcine gut remain unclear. In this study, multi-omics analysis was employed in a Yuxi black pig animal model to explore the possible regulatory mechanism of mulberry leaves on pork quality. Sixty Yuxi black pigs were divided into two groups: the control group (n = 15) was fed a standard diet, and the experimental group (n = 45) was fed a diet supplemented with 8% mulberry leaves. Experiments were performed in three replicates (n = 15 per replicate); the two diets were ensured to be nutritionally balanced, and the feeding period was 120 days. The results showed that pigs receiving the diet supplemented with mulberry leaves had significantly reduced backfat thickness (p < 0.05) and increased intramuscular fat (IMF) content (p < 0.05) compared with pigs receiving the standard diet. Lipidomics analysis showed that mulberry leaves improved the lipid profile composition and increased the proportion of triglycerides (TGs). Interestingly, the IMF content was positively correlated with acyl C18:2 and negatively correlated with C18:1 of differential TGs. In addition, the cecal microbiological analysis showed that mulberry leaves could increase the abundance of bacteria such as UCG-005, Muribaculaceae_norank, Prevotellaceae_NK3B31_group, and Limosilactobacillus. Simultaneously, the relative levels of L-tyrosine-ethyl ester, oleic acid methyl ester, 21-deoxycortisol, N-acetyldihydrosphingosine, and mulberrin were increased. Furthermore, we found that mulberry leaf supplementation significantly increased the mRNA expression of lipoprotein lipase, fatty acid-binding protein 4, and peroxisome proliferators-activated receptor γ in muscle (p < 0.01). Mulberry leaf supplementation significantly increased the mRNA expression of diacylglycerol acyltransferase 1 (p < 0.05) while significantly decreasing the expression of acetyl CoA carboxylase in backfat (p < 0.05). Furthermore, mulberry leaf supplementation significantly upregulated the mRNA expression of hormone-sensitive triglyceride lipase and peroxisome proliferator-activated receptor α (p < 0.05) in backfat. In addition, mulberry leaf supplementation led to increased serum leptin and adiponectin (p < 0.01). Collectively, this omic profile is consistent with an increased ratio of IMF to backfat in the pig model.

Technological Tools and Artificial Intelligence in Estrus Detection of Sows-A Comprehensive Review

In animal farming, timely estrus detection and prediction of the best moment for insemination is crucial. Traditional sow estrus detection depends on the expertise of a farm attendant which can be inconsistent, time-consuming, and labor-intensive. Attempts and trials in developing and implementing technological tools to detect estrus have been explored by researchers. The objective of this review is to assess the automatic methods of estrus recognition in operation for sows and point out their strong and weak points to assist in developing new and improved detection systems. Real-time methods using body and vulvar temperature, posture recognition, and activity measurements show higher precision. Incorporating artificial intelligence with multiple estrus-related parameters is expected to enhance accuracy. Further development of new systems relies mostly upon the improved algorithm and accurate data provided. Future systems should be designed to minimize the misclassification rate, so better detection is achieved.

Pig Movement Estimation by Integrating Optical Flow with a Multi-Object Tracking Model

Pig husbandry constitutes a significant segment within the broader framework of livestock farming, with porcine well-being emerging as a paramount concern due to its direct implications on pig breeding and production. An easily observable proxy for assessing the health of pigs lies in their daily patterns of movement. The daily movement patterns of pigs can be used as an indicator of their health, in which more active pigs are usually healthier than those who are not active, providing farmers with knowledge of identifying pigs' health state before they become sick or their condition becomes life-threatening. However, the conventional means of estimating pig mobility largely rely on manual observations by farmers, which is impractical in the context of contemporary centralized and extensive pig farming operations. In response to these challenges, multi-object tracking and pig behavior methods are adopted to monitor pig health and welfare closely. Regrettably, these existing methods frequently fall short of providing precise and quantified measurements of movement distance, thereby yielding a rudimentary metric for assessing pig health. This paper proposes a novel approach that integrates optical flow and a multi-object tracking algorithm to more accurately gauge pig movement based on both qualitative and quantitative analyses of the shortcomings of solely relying on tracking algorithms. The optical flow records accurate movement between two consecutive frames and the multi-object tracking algorithm offers individual tracks for each pig. By combining optical flow and the tracking algorithm, our approach can accurately estimate each pig's movement. Moreover, the incorporation of optical flow affords the capacity to discern partial movements, such as instances where only the pig's head is in motion while the remainder of its body remains stationary. The experimental results show that the proposed method has superiority over the method of solely using tracking results, i.e., bounding boxes. The reason is that the movement calculated based on bounding boxes is easily affected by the size fluctuation while the optical flow data can avoid these drawbacks and even provide more fine-grained motion information. The virtues inherent in the proposed method culminate in the provision of more accurate and comprehensive information, thus enhancing the efficacy of decision-making and management processes within the realm of pig farming.

Individual Pig Identification Using Back Surface Point Clouds in 3D Vision

The individual identification of pigs is the basis for precision livestock farming (PLF), which can provide prerequisites for personalized feeding, disease monitoring, growth condition monitoring and behavior identification. Pig face recognition has the problem that pig face samples are difficult to collect and images are easily affected by the environment and body dirt. Due to this problem, we proposed a method for individual pig identification using three-dimension (3D) point clouds of the pig's back surface. Firstly, a point cloud segmentation model based on the PointNet++ algorithm is established to segment the pig's back point clouds from the complex background and use it as the input for individual recognition. Then, an individual pig recognition model based on the improved PointNet++LGG algorithm was constructed by increasing the adaptive global sampling radius, deepening the network structure and increasing the number of features to extract higher-dimensional features for accurate recognition of different individuals with similar body sizes. In total, 10,574 3D point cloud images of ten pigs were collected to construct the dataset. The experimental results showed that the accuracy of the individual pig identification model based on the PointNet++LGG algorithm reached 95.26%, which was 2.18%, 16.76% and 17.19% higher compared with the PointNet model, PointNet++SSG model and MSG model, respectively. Individual pig identification based on 3D point clouds of the back surface is effective. This approach is easy to integrate with functions such as body condition assessment and behavior recognition, and is conducive to the development of precision livestock farming.

Double-Camera Fusion System for Animal-Position Awareness in Farming Pens

In livestock breeding, continuous and objective monitoring of animals is manually unfeasible due to the large scale of breeding and expensive labour. Computer vision technology can generate accurate and real-time individual animal or animal group information from video surveillance. However, the frequent occlusion between animals and changes in appearance features caused by varying lighting conditions makes single-camera systems less attractive. We propose a double-camera system and image registration algorithms to spatially fuse the information from different viewpoints to solve these issues. This paper presents a deformable learning-based registration framework, where the input image pairs are initially linearly pre-registered. Then, an unsupervised convolutional neural network is employed to fit the mapping from one view to another, using a large number of unlabelled samples for training. The learned parameters are then used in a semi-supervised network and fine-tuned with a small number of manually annotated landmarks. The actual pixel displacement error is introduced as a complement to an image similarity measure. The performance of the proposed fine-tuned method is evaluated on real farming datasets and demonstrates significant improvement in lowering the registration errors than commonly used feature-based and intensity-based methods. This approach also reduces the registration time of an unseen image pair to less than 0.5 s. The proposed method provides a high-quality reference processing step for improving subsequent tasks such as multi-object tracking and behaviour recognition of animals for further analysis.

The Research Progress of Vision-Based Artificial Intelligence in Smart Pig Farming

Pork accounts for an important proportion of livestock products. For pig farming, a lot of manpower, material resources and time are required to monitor pig health and welfare. As the number of pigs in farming increases, the continued use of traditional monitoring methods may cause stress and harm to pigs and farmers and affect pig health and welfare as well as farming economic output. In addition, the application of artificial intelligence has become a core part of smart pig farming. The precision pig farming system uses sensors such as cameras and radio frequency identification to monitor biometric information such as pig sound and pig behavior in real-time and convert them into key indicators of pig health and welfare. By analyzing the key indicators, problems in pig health and welfare can be detected early, and timely intervention and treatment can be provided, which helps to improve the production and economic efficiency of pig farming. This paper studies more than 150 papers on precision pig farming and summarizes and evaluates the application of artificial intelligence technologies to pig detection, tracking, behavior recognition and sound recognition. Finally, we summarize and discuss the opportunities and challenges of precision pig farming.