Search

Posture transformation characteristics of sows in late gestation can be used as intuitive indicators to predict sow delivery time. To reduce the imbalance in posture distribution and insignificant differences between coherent sow postures in late gestation, a model based on Libra R-CNN was proposed to recognise the sow postures. Libra R-CNN is a simple but effective framework that incorporates intersection over union (IoU)-balanced sampling, a balanced feature pyramid, and balanced L1 loss, aiming to balance learning for object detection. The model used here realised the recognition of sow postures: lateral, sternum, sitting, and standing. Experimental results revealed close relationships between the two basic characteristics ('ratio of express nesting behaviour' and 'frequency of postural transformation') and delivery time. The mean average precision values for IoU = 0.5 (mAP 0.5) of Libra R-CNN on the validation and test sets from the sow posture sample library were 0.97 and 0.96, whilst the AP value of various postures was balanced. From 16 to 4 h before sow farrowing, the values of the two basic characteristics were significantly increased and lasted for a long time, and started to decrease rapidly 4 h before farrowing. Heat stress caused sow agitation but decreased nesting behaviour before parturition. The Pearson correlation coefficients of 15 features constructed based on the two characteristics and delivery time were all above 0.6. Therefore, Libra R-CNN realised the detection of sow posture and the two basic characteristics can be used as effective indicators for the automated prediction of sow delivery time. • Nesting behaviour and the frequency of postural transformation investigated. • Values of both characteristics significantly increased before sows farrowing. • Characteristics used as effective indicators for sow delivery prediction. • The mAP 0.5 of Libra R-CNN for the validation sets was 0.97. • Heat stress shown to cause sow agitation, but decreased nesting behaviour. [ABSTRACT FROM AUTHOR]

In this paper, a novel pig detection approach is proposed for video surveillance based on overhead views in complex group-housed environments. In contrast to traditional approaches that use basic image processing techniques, the proposed algorithm focuses on particular appearance features of pigs and employs the template matching framework to facilitate pig detection. First, to capture the key appearance characteristics of pigs in templates, we represent images as the combination of two complementary descriptors, i.e., dominant orientation templates (DOTs) and brightness ratio templates (BRTs). A DOT establishes relationships with edge features, while a BRT encodes intensity and texture features. Both descriptors are given in binary form to increase the time efficiency of the algorithm. Then, representative pig templates can be obtained from training images through an automatic selection process based on template clustering. Finally, the input image is scanned to compute matching responses for all templates. Thus, objects having similar visual properties as the given templates are considered as detected pigs. Experiments are conducted on the test data extracted from a surveillance video covering 5 days. The experimental results show that our algorithm achieves an average detection rate of 86.8% with a low false alarm rate, which outperforms methods based on threshold segmentation. As a result, the proposed method can effectively and reliably extract pigs in complex scenes including multiple sources of disturbances such as uneven illumination, foreground objects of varying colour, pigs moving slowly, etc. • A novel pig detection approach using template matching is presented. • Multiple visual features are combined in templates to describe appearances of pigs. • Representative templates are automatically selected through clustering algorithm. • Compared with threshold-segmentation-based methods, better performance is achieved in natural group-housed environments. [ABSTRACT FROM AUTHOR]

Abstract This paper establishes two logarithmically improved regularity criteria for the Cauchy problem of the 3D Boussinesq equations in termsof the pressure on framework of the anisotropic Lebesgue spaces. More precisely, it is proved that if ∫ 0 T ‖ ‖ ‖ P ‖ L x 1 p ‖ L x 2 q ‖ L x 3 r β 1 + ln (1 + ‖ θ (⋅ , t) ‖ L 4 ) d t < ∞ , where 2 β + 1 p + 1 q + 1 r = 2 and 2 ≤ p , q , r ≤ ∞ , 1 − (1 p + 1 q + 1 r) ≥ 0 , or ∫ 0 T ‖ ‖ ‖ ∇ P ‖ L x 1 p ‖ L x 2 q ‖ L x 3 r β 1 + ln (1 + ‖ θ (⋅ , t) ‖ L 4 ) d t < ∞ , where 2 β + 1 p + 1 q + 1 r = 3 and 1 ≤ p , q , r ≤ ∞ , 1 − (1 2 p + 1 2 q + 1 2 r) ≥ 0 , for some T > 0 , then the corresponding weak solution (u , θ) to the 3D Boussinesq equations is regular on [ 0 , T ]. [ABSTRACT FROM AUTHOR]

The portability and stability of algorithms from hyperspectral imaging (HIS) to multispectral imaging (MIS) for detection of common defects on fruits is limiting the commercialization of multispectral inspection systems. The efficiency of selected wavelengths, the robustness to physical and biological variability, and the transplantation of the classification algorithms affect the portability and stability of algorithms from HIS to MIS. Apples with wind damage, insect damage, bruises, decay, hail, russeting, spot, scar, stem, calyx, as well as with sound surfaces were studied. Different image recognition methods were proposed according to the defect type and spectral characters. For common defects, the overall detection accuracy of 93.6% for HIS data and 91.4% for MIS with the same wavelengths and same methods was obtained. The uneven lightness distribution in the spectral images was also investigated and eliminated by the band math or partial least squares (PLS) model based methods Principal component analysis, minimum noise fraction, band math, particle least square model based methods were also investigated and compared. Results show that band math and model based method possess a higher portability and stability from HIS to MIS, and the proposed methods could make the variation of illumination, physical and biological variability as negligible interference in common defect detection. This study could provide potential reference for promoting the development of MIS and the applications of transplantable algorithms in HIS. [ABSTRACT FROM AUTHOR]

• A local outlier factor algorithm detects social density of suckling piglets. • This method could quantify the spatial relationship among different piglets. • This method could detect cluster in the piglet group. • The proposed method could detect outlier piglets. Pigs and their lactating piglets are herd animals, and breeders usually observe their social distance to determine their physiological status, such as when they form clusters during cold stress, or whether there are outlier piglets with abnormal growth conditions. Instead, papers on automated detection of piglet social distance are rare. This paper proposed a novel method, the convolutional neural network combined with modified local outlier factor (CNN-LOF), to quantify the piglet social density, and detect piglets far from the herd (outlier piglets). The convolutional neural network (CNN) model named YOLOv5 was used to construct the piglet detector, and auto-mark piglets by detection box, which adopted boxes' center points instead of the piglets. The optimized local outlier factor (LOF) algorithm was employed to calculate the social density of the piglets, and the outlier piglets were based on outlier factors greater than 2. Besides, the social density of different periods was calculated, compared, and analyzed. According to the results, the accuracy of CNN-LOF for detecting outlier piglets was 97.7% for the 6,113 test images measured. The larger social density of a piglet indicates more other piglets around and higher probability of being in the center of the herd. CNN-LOF and manual detection have a similarity of more than 80% in a continuous period of 1 h. In summary, this study quantifies the social density of suckling piglets, which also intuitively reveals the distribution of piglets in different environments, and provides technical support for precision livestock farming. [ABSTRACT FROM AUTHOR]

To establish a healthy model of sow and analyze temperature variation of childbirth and estrus process, a remote and real-time monitoring temperature node based on MLX90614 and CC2430 was designed for sows, which can provide a scientific basis for disease diagnosis and estrus judgement. CC2430 MCU was the master controller of the node and the temperature of sows was measured by MLX90614 which was a high-precision infrared temperature sensor. CC2430 single-chip can communicate with MLX90614 and the data was read and transmitted to the gateway wireless node through SMbus protocol. In the testing process, sow's buttock as the best measurement area, the correlation of the sow's buttocks and rectum was studied, and result of the temperature was compensated in order to increase the accuracy. The experiment showed that the effective communication distance of sensor node was thirty meters, and the system of sows temperature measurement error was 0.21 °C, which could meet the requirement of animal livestock and poultry breeding industry. A non-contact method of accurate and real-time measurement of the temperature of sows was innovatively proposed, providing technical support for large-scale scientific breeding of sows. [ABSTRACT FROM AUTHOR]

Sow farrowing is an important part of pig breeding. The accurate and effective early warning of sow behaviors in farrowing helps breeders determine whether it is necessary to intervene with the farrowing process in a timely manner and is thus essential for increasing the survival rate of piglets and the profits of pig farms. For large pig farms, human resources and costs are important considerations in farrowing supervision. The existing method, which uses cloud computing-based deep learning to supervise sow farrowing, has a high equipment cost and requires uploading all data to a cloud data center, requiring a large network bandwidth. Thus, this paper proposes an approach for the early warning and supervision of farrowing behaviors based on the embedded artificial-intelligence computing platform (NVIDIA Jetson Nano). This lightweight deep learning method allows the rapid processing of sow farrowing video data at edge nodes, reducing the bandwidth requirement and ensuring data security in the network transmission. Experiments indicated that after the model was migrated to the Jetson Nano, its precision of sow postures and newborn piglets detection was 93.5%, with a recall rate of 92.2%, and the detection speed was increased by a factor larger than 8. The early warning of 18 approaching farrowing (5 h) sows were tested. The mean error of warning was 1.02 h. [ABSTRACT FROM AUTHOR]

This study aimed to investigate the effects of heat stress on posture transitions of perinatal primiparous sows around the parturition period from 72 h prepartum to 24 h postpartum. The reproductive performance of sows was measured, and the relationship between posture transitions and reproductive performance was also analyzed. Ten primiparous sows were randomly assigned to thermoneutral (TN) (18–22 °C; n = 5) or heat stress (HS) (28–32 °C; n = 5) treatments. Posture transitioning, including the frequency of posture change, duration of dynamic posture (DP), and lateral lying with udder to the piglet creep box (PCB) during three periods (72 h prepartum, sub-partum, and 24 h postpartum, respectively), were recorded. Posture change frequency was significantly increased, starting from 24 h prepartum to the onset of farrowing in both the TN (P < 0.05) and HS (P < 0.01) groups. Moreover, the peak value of posture change frequency in the TN group was concentrated during the 12 h prepartum period, positively correlated with the quantities of head-first birth piglets and sub-partum duration, respectively. DP duration increased during the period of 24 h prepartum and then decreased sharply (P < 0.001 and P < 0.05 for TN and HS groups, respectively). The duration of facing the udder to the PCB increased during sub-partum and postpartum TN (P < 0.001). The duration of sub-partum (P < 0.05) and delivery time of single piglets (P < 0.01) in the HS group was prolonged, and piglets from the HS group had a lower weight gain than the TN group both at d10 (P < 0.001) and weaning time (P < 0.001). In conclusion, HS had obvious adverse effects on nursery behavior and reproductive abilities in perinatal primiparous sows, which resulted in poor growth performance of lactating piglets. • The effect of heat stress on posture transitions of perinatal sows were investigated. • The delivery time of single piglet affected by heat stress. • The effect of heat stress on reproductive performance of sows were investigated. • Sows' posture transitions were related to the reproductive performance. [ABSTRACT FROM AUTHOR]

Simple Summary: The utilization of computer vision technology and automation for monitoring and collecting pigeon eggs is of significant importance for improving labor productivity and the breeding of egg-producing pigeons. Currently, research both domestically and internationally has predominantly focused on the detection of eggs from poultry such as chickens, ducks, and geese, leaving pigeon egg recognition largely unexplored. This study proposes an effective and lightweight network model, YOLOv8-PG, based on YOLOv8n, which maintains high detection accuracy while reducing the model's parameter count and computational load. This approach facilitates cost reduction in deployment and enhances feasibility for implementation on mobile robotic platforms. In response to the high breakage rate of pigeon eggs and the significant labor costs associated with egg-producing pigeon farming, this study proposes an improved YOLOv8-PG (real versus fake pigeon egg detection) model based on YOLOv8n. Specifically, the Bottleneck in the C2f module of the YOLOv8n backbone network and neck network are replaced with Fasternet-EMA Block and Fasternet Block, respectively. The Fasternet Block is designed based on PConv (Partial Convolution) to reduce model parameter count and computational load efficiently. Furthermore, the incorporation of the EMA (Efficient Multi-scale Attention) mechanism helps mitigate interference from complex environments on pigeon-egg feature-extraction capabilities. Additionally, Dysample, an ultra-lightweight and effective upsampler, is introduced into the neck network to further enhance performance with lower computational overhead. Finally, the EXPMA (exponential moving average) concept is employed to optimize the SlideLoss and propose the EMASlideLoss classification loss function, addressing the issue of imbalanced data samples and enhancing the model's robustness. The experimental results showed that the F1-score, mAP50-95, and mAP75 of YOLOv8-PG increased by 0.76%, 1.56%, and 4.45%, respectively, compared with the baseline YOLOv8n model. Moreover, the model's parameter count and computational load are reduced by 24.69% and 22.89%, respectively. Compared to detection models such as Faster R-CNN, YOLOv5s, YOLOv7, and YOLOv8s, YOLOv8-PG exhibits superior performance. Additionally, the reduction in parameter count and computational load contributes to lowering the model deployment costs and facilitates its implementation on mobile robotic platforms. [ABSTRACT FROM AUTHOR]

Simple Summary: Pig farming plays a critical role in global animal husbandry, with pig weight and body dimensions serving as key indicators of growth and development. Manual measurement methods currently used face challenges like herding difficulties and stress in pigs. To address this, this study introduces a non-contact weight estimation and body measurement model using point cloud data from pig backs. By utilizing a depth camera and advanced algorithms, accurate weight predictions and body size measurements were achieved. The model showed promising results, with reduced errors compared to manual methods, demonstrating its potential for improving efficiency and animal welfare in pig farming practices. Pig farming is a crucial sector in global animal husbandry. The weight and body dimension data of pigs reflect their growth and development status, serving as vital metrics for assessing their progress. Presently, pig weight and body dimensions are predominantly measured manually, which poses challenges such as difficulties in herding, stress responses in pigs, and the control of zoonotic diseases. To address these issues, this study proposes a non-contact weight estimation and body measurement model based on point cloud data from pig backs. A depth camera was installed above a weighbridge to acquire 3D point cloud data from 258 Yorkshire–Landrace crossbred sows. We selected 200 Yorkshire–Landrace sows as the research subjects and applied point cloud filtering and denoising techniques to their three-dimensional point cloud data. Subsequently, a K-means clustering segmentation algorithm was employed to extract the point cloud corresponding to the pigs' backs. A convolutional neural network with a multi-head attention was established for pig weight prediction and added RGB information as an additional feature. During the data processing process, we also measured the back body size information of the pigs. During the model evaluation, 58 Yorkshire–Landrace sows were specifically selected for experimental assessment. Compared to manual measurements, the weight estimation exhibited an average absolute error of 11.552 kg, average relative error of 4.812%, and root mean square error of 11.181 kg. Specifically, for the MACNN, incorporating RGB information as an additional feature resulted in a decrease of 2.469 kg in the RMSE, a decrease of 0.8% in the MAPE, and a decrease of 1.032 kg in the MAE. Measurements of shoulder width, abdominal width, and hip width yielded corresponding average relative errors of 3.144%, 3.798%, and 3.820%. In conclusion, a convolutional neural network with a multi-head attention was established for pig weight prediction, and incorporating RGB information as an additional feature method demonstrated accuracy and reliability for weight estimation and body dimension measurement. [ABSTRACT FROM AUTHOR]

The occurrence of poultry diseases not only affects farm production economics but also leads to poor poultry welfare, food safety concerns, and zoonotic infections. Therefore, timely detection of these diseases is of paramount importance in poultry production. This study proposes a machine vision-based monitoring system for broiler chicken as they walk through a test area. Data were collected from two groups of broilers; control group and treatment group (inoculated intramuscularly with virulent Newcastle disease virus) housed in fully isolated chambers for comparative monitoring. The broilers were monitored by video surveillance for data labelling and depth camera for the automated health status classifier development. Feature variables were extracted based on 2D posture shape descriptors (circle variance, elongation, convexity, complexity, and eccentricity) and mobility feature (walk speed). A statistical analysis of the feature variables established that all investigated features were statistically significant (p < 0.05) with time after challenge in the treatment group. The earliest possible infection detection time was on the 4th day based on circle variance and elongation, and the 6th day based on eccentricity and walk speed. However, convexity and complexity could not provide early detection. Two sets of classifiers were then developed based on only the posture shape descriptors, and on all the feature variables, The Support Vector Machine (RBF-SVM) outperformed all the other models with an accuracy of 0.975 and 0.978 respectively. The proposed system can serve as an automatic broiler monitoring system by providing an early warning and prediction of an occurrence of disease continuously and non-intrusively. • Machine vision system developed to non-intrusively monitor broiler chicken. • 2D posture shape descriptors and walk speed were used as the feature variables. • Achieved early infection detection based on the extracted feature variables. • The RBF-SVM was the best model at an accuracy of 0.978 on the testing set. [ABSTRACT FROM AUTHOR]

Sow body condition scoring has been confirmed as a vital procedure in sow management. A timely and accurate assessment of the body condition of a sow is conducive to determining nutritional supply, and it takes on critical significance in enhancing sow reproductive performance. Manual sow body condition scoring methods have been extensively employed in large-scale sow farms, which are time-consuming and labor-intensive. To address the above-mentioned problem, a dual neural network-based automatic scoring method was developed in this study for sow body condition. The developed method aims to enhance the ability to capture local features and global information in sow images by combining CNN and transformer networks. Moreover, it introduces a CBAM module to help the network pay more attention to crucial feature channels while suppressing attention to irrelevant channels. To tackle the problem of imbalanced categories and mislabeling of body condition data, the original loss function was substituted with the optimized focal loss function. As indicated by the model test, the sow body condition classification achieved an average precision of 91.06%, the average recall rate was 91.58%, and the average F1 score reached 91.31%. The comprehensive comparative experimental results suggested that the proposed method yielded optimal performance on this dataset. The method developed in this study is capable of achieving automatic scoring of sow body condition, and it shows broad and promising applications. [ABSTRACT FROM AUTHOR]

In aquaculture, accurately counting fish fries is a prerequisite for fish population management and marketing. However, due to the fry overlapping and occlusion issues, the manual counting method is time-consuming, and the counting result might be inaccurate. Therefore, we utilized the computer vision technique to develop a lightweight fish fry counting model to address this issue. First, we constructed a diverse dataset containing images of largemouth bass in varying numbers, captured under different lighting conditions and water depths. Then, we optimized the YOLOv5s model by channel pruning to reduce the model size, while maintaining the detection accuracy. Through extensive experiments, we examined the effect of different pruning rates on the model performance and compared the pruned YOLOv5s model with state-of-the-art detection models through the evaluation criteria like precision, recall, mean average precision (mAP), model size, GFLOPs, and detection speed. In addition, we investigated the impact of environmental factors, such as lighting conditions and water depths, on the detection performance of the pruned YOLOv5s model. The experimental results demonstrated that the YOLOv5s model with a pruning rate of 15% achieved over 90% accuracy and 13 FPS in the dense and complex scenes, which met the practical requirement for the fry counting task. In addition, we also identified that the pruned YOLOv5s model would achieve the optimal performance under the white illumination and shallow water depth setting. In conclusion, this study provided an efficient solution for fish fry counting, and the proposed model was expected to be applied in the real-world application. [ABSTRACT FROM AUTHOR]

Highlights • A new algorithm to extract piglet ear base temperature by using thermal image processing. • Identification of piglet head part by using SVM based on contour features. • Ear base points are located based on the shape features of piglet head part. • More than 97% of the testing images whose ear base temperatures can be obtained with the error less than 0.4 °C. Abstract Ear bases are considered the thermal windows of a piglet. Temperature variation in piglet ear bases can be used as the indicator of a piglet's health status. However, piglet skin temperatures in thermal windows in the existing research are obtained manually from infrared thermal images captured by a thermography. This has put an obstacle at the automatic identification of piglets with health disorder. An algorithm was proposed in this paper to extract ear base temperature automatically from top view piglet thermal images. Firstly, a SVM (Support Vector Machine) classifier was trained to identify piglet head part. Then, two ear base points were located based on the shape feature of the head part contour. Finally, two maximum temperatures inside the two circles centered by ear base points were extracted as the ear base temperatures. The proposed algorithm was implemented in Matlab® (R2016a) and applied to 100 testing images. The extracted ear base temperatures were compared with those extracted manually by using Fluke SmartView 3.14 (FLUKE Systems). Comparison results showed that for left and right ear base respectively, 97% and 98% of the testing images had an error within 0.4 °C. Ear base temperatures with such accuracy provided a foundation for the automatic identification of sick piglets. [ABSTRACT FROM AUTHOR]

The average weight of piglets in lactation can be monitored automatically by piglets’ average weight monitoring systems which are designed based on wireless multimedia sensor networks. Piglets counting in an automatic manner for piglets images is the foundation of these systems. Adhesive piglets may exist in an image due to the social character of piglets, which challenges the image splitting and automatic piglets counting. This paper proposes a segmentation algorithm for adhesive piglets images based on ellipse fitting method. Firstly, ellipse fitting is implemented for a large number of images which have one piglet. Parameters range of ellipses fitted by images with a single piglet on different age is extracted. Secondly, contours of connected components in an adhesive piglets image are extracted. Each contour is segmented based on concave points. Ellipse fitting is implemented for each contour segment. Finally, 5 rules for ellipse merging are put forwarded, which are used to merge anomalous ellipses. After ellipse merging, the number of ellipses equals the number of piglets. The proposed algorithm is applied to adhesive piglets images in Matlab R2012b and the experimental results show that the counting accuracy exceeds 86% when the number of piglets is less than 7. The algorithm provides the foundation for the piglets’ average weight monitoring systems. [ABSTRACT FROM AUTHOR]

To acquire greenhouse enviromental information,besides the acquisition of greenhouse enviromental information such as temperature and humidity, light, carbon dioxide and other enviromental information, and the image information of plants growing condition is also very important. Aim at solving the requirements of plants growing information in greenhouse enviroment and the problems that existing node transmission about ZigBee protocol stack is complex and high power consumption,this study designed a wireless smart image acquisition node in order to monitor the plant growth in greenhouse,and introduced a sleep-wake up mechanism to reduce the energy of cosumption. The node is composed by STM32ZET6 chip as micro-controller,CC1101 for RF module,and expanded OV7670 as image senor. Image package was designed according to the frame format of the SimpliciTI protocol,and after JPEG2000 compression subcontrated and transmissed by network,the server recieved the image and using the improved LoG operator achieved image segmentation,and obtained the image segmentation results. The results privide image analysis data for greenhouse enviroment monitoting. To test the relation between transmitted power with data transmission rate and the package loss rate. RF working frequency is 433 MHz,and adopts the GFSK modulation method. Antenna height distance is 1 m,antenna gain 2 dBi is chosen,the sending power respectively is 250 kBaud. The length of test package is 250 kBaud,to send 100 package each time. The experiments showed the design of wireless sensor node can collect and transmit plant growth information at a relatively low energy consumption,the maximum communication distance is 110 m and the loss ration of packet is lower. An image at interval of 20 mins, and 21 images together were collected. PC got cucumber leaf images through the iamge collecting software designed through VC++6.0,selected the Sobel operator, LoG operator and modified LoG operator. A frame of JPEG2000 format upon plant growth in greenhouse information for the image size is about 18 kB,and transmission time is shorter. Nodes under 3.6 V lithium battery can work 20 days, and gathering node receives a frame JPEG2000 compressed image upon time consumption of 158 s. The results show the segmentation method is robust, division of larger changes in the light of the circumstances remains stable, and it improves the bigger influence of noise, over-detailed division status. [ABSTRACT FROM AUTHOR]