Acquiring phenotypic data from livestock constitutes a crucial yet cumbersome phase in the breeding process. Traditionally, obtaining livestock phenotypic data primarily involves manual, on-body measurement methods. This approach not only requires extensive labor but also induces stress on animals, which leads to potential economic losses. Presently, the integration of next-generation Artificial Intelligence (AI), visual processing, intelligent sensing, multimodal fusion processing, and robotic technology is increasingly prevalent in livestock farming. The advantages of these technologies lie in their rapidity and efficiency, coupled with their capability to acquire livestock data in a non-contact manner. Based on this, we provide a comprehensive summary and analysis of the primary advanced technologies employed in the non-contact acquisition of livestock phenotypic data. This review focuses on visual and AI-related techniques, including 3D reconstruction technology, body dimension acquisition techniques, and live animal weight estimation. We introduce the development of livestock 3D reconstruction technology and compare the methods of obtaining 3D point cloud data of livestock through RGB cameras, laser scanning, and 3D cameras. Subsequently, we explore body size calculation methods and compare the advantages and disadvantages of RGB image calculation methods and 3D point cloud body size calculation methods. Furthermore, we also compare and analyze weight estimation methods of linear regression and neural networks. Finally, we discuss the challenges and future trends of non-contact livestock phenotypic data acquisition. Through emerging technologies like next-generation AI and computer vision, the acquisition, analysis, and management of livestock phenotypic data are poised for rapid advancement.

Accurate monitoring of animal behaviors enables improved management in precision livestock farming (PLF), supporting critical applications including health assessment, estrus detection, parturition monitoring, and feed intake estimation. Although both contact and non-contact sensing modalities are utilized, wearable devices with embedded sensors (e.g., accelerometers, pressure sensors) offer unique advantages through continuous data streams that enhance behavioral traceability. Focusing specifically on contact sensing techniques, this review examines sensor characteristics and data acquisition challenges, methodologies for processing behavioral data and implementing identification algorithms, industrial applications enabled by recognition outcomes, and prevailing challenges with emerging research opportunities. Current behavior classification relies predominantly on traditional machine learning or deep learning approaches with high-frequency data acquisition. The fundamental limitation restricting advancement in this field is the difficulty in maintaining high-fidelity recognition performance at reduced acquisition rates, particularly for integrated multi-behavior identification. Considering that the computational demands and limited adaptability to complex field environments remain significant constraints, Tiny Machine Learning (Tiny ML) could present opportunities to guide future research toward practical, scalable behavioral monitoring solutions. In addition, algorithm development for functional applications post behavior recognition may represent a critical future research direction.

Natural breeding scenes have the characteristics of a large number of cows, complex lighting, and a complex background environment, which presents great difficulties for the detection of dairy cow estrus behavior. However, the existing research on cow estrus behavior detection works well in ideal environments with a small number of cows and has a low inference speed and accuracy in natural scenes. To improve the inference speed and accuracy of cow estrus behavior in natural scenes, this paper proposes a cow estrus behavior detection method based on the improved YOLOv5. By improving the YOLOv5 model, it has stronger detection ability for complex environments and multi-scale objects. First, the atrous spatial pyramid pooling (ASPP) module is employed to optimize the YOLOv5l network at multiple scales, which improves the model’s receptive field and ability to perceive global contextual multiscale information. Second, a cow estrus behavior detection model is constructed by combining the channel-attention mechanism and a deep-asymmetric-bottleneck module. Last, K-means clustering is performed to obtain new anchors and complete intersection over union (CIoU) is used to introduce the relative ratio between the predicted box of the cow mounting and the true box of the cow mounting to the regression box prediction function to improve the scale invariance of the model. Multiple cameras were installed in a natural breeding scene containing 200 cows to capture videos of cows mounting. A total of 2668 images were obtained from 115 videos of cow mounting events from the training set, and 675 images were obtained from 29 videos of cow mounting events from the test set. The training set is augmented by the mosaic method to increase the diversity of the dataset. The experimental results show that the average accuracy of the improved model was 94.3%, that the precision was 97.0%, and that the recall was 89.5%, which were higher than those of mainstream models such as YOLOv5, YOLOv3, and Faster R-CNN. The results of the ablation experiments show that ASPP, new anchors, C3SAB, and C3DAB designed in this study can improve the accuracy of the model by 5.9%. Furthermore, when the ASPP dilated convolution was set to (1,5,9,13) and the loss function was set to CIoU, the model had the highest accuracy. The class activation map function was utilized to visualize the model’s feature extraction results and to explain the model’s region of interest for cow images in natural scenes, which demonstrates the effectiveness of the model. Therefore, the model proposed in this study can improve the accuracy of the model for detecting cow estrus events. Additionally, the model’s inference speed was 71 frames per second (fps), which meets the requirements of fast and accurate detection of cow estrus events in natural scenes and all-weather conditions.

The use of an accelerometer is considered as a promising method for the automatic measurement of the feeding behavior or feed intake of cattle, with great significance in facilitating daily management. To address further need for commercial use, an efficient classification algorithm at a low sample frequency is needed to reduce the amount of recorded data to increase the battery life of the monitoring device, and a high-precision model needs to be developed to predict feed intake on the basis of feeding behavior. Accelerograms for the jaw movement and feed intake of 13 mid-lactating cows were collected during feeding with a sampling frequency of 1 Hz at three different positions: the nasolabial levator muscle (P1), the right masseter muscle (P2), and the left lower lip muscle (P3). A behavior identification framework was developed to recognize jaw movements including ingesting, chewing and ingesting–chewing through extreme gradient boosting (XGB) integrated with the hidden Markov model solved by the Viterbi algorithm (HMM–Viterbi). Fourteen machine learning models were established and compared in order to predict feed intake rate through the accelerometer signals of recognized jaw movement activities. The developed behavior identification framework could effectively recognize different jaw movement activities with a precision of 99% at a window size of 10 s. The measured feed intake rate was 190 ± 89 g/min and could be predicted efficiently using the extra trees regressor (ETR), whose R², RMSE, and NME were 0.97, 0.36 and 0.05, respectively. The three investigated monitoring sites may have affected the accuracy of feed intake prediction, but not behavior identification. P1 was recommended as the proper monitoring site, and the results of this study provide a reference for the further development of a wearable device equipped with accelerometers to measure feeding behavior and to predict feed intake.

The physiological well-being of dairy cows is intimately tied to their behavior. Detecting aberrant dairy cows early and reducing financial losses on farms are both possible with real-time and reliable monitoring of their behavior. The behavior data of dairy cows in real environments have dense occlusion and multi-scale issues, which affect the detection results of the model. Therefore, we focus on both data processing and model construction to improve the results of dairy cow behavior detection. We use a mixed data augmentation method to provide the model with rich cow behavior features. Simultaneously refining the model to optimize the detection outcomes of dairy cow behavior amidst challenging conditions, such as dense occlusion and varying scales. First, a Res2 backbone was constructed to incorporate multi-scale receptive fields and improve the YOLOv3’s backbone for the multi-scale feature of dairy cow behaviors. In addition, YOLOv3 detectors were optimized to accurately locate individual dairy cows in different dense environments by combining the global location information of images, and the Global Context Predict Head was designed to enhance the performance of recognizing dairy cow behaviors in crowded surroundings. The dairy cow behavior detection model we built has an accuracy of 90.6%, 91.7%, 80.7%, and 98.5% for the four behaviors of dairy cows standing, lying, walking, and mounting, respectively. The average accuracy of dairy cow detection is 90.4%, which is 1.2% and 12.9% higher than the detection results of YOLOV3, YOLO-tiny and other models respectively. In comparison to YOLOv3, the Average Precision evaluation of the model improves by 2.6% and 1.4% for two similar features of walking and standing behaviors, respectively. The recognition results prove that the model generalizes better for recognizing dairy cow behaviors using behavior videos in various scenes with multi-scale and dense environment features.

As machine vision technology has advanced, pig face recognition has gained wide attention as an individual pig identification method. This study establishes an improved ResNAM network as a backbone network for pig face image feature extraction by combining an NAM (normalization-based attention module) attention mechanism and a ResNet model to probe non-contact open-set pig face recognition. Then, an open-set pig face recognition framework is designed by integrating three loss functions and two metrics to finish the task with no crossover of individuals in the training and test sets. The SphereFace loss function with the cosine distance as a metric and ResNAM are combined in the framework to obtain the optimal open-set pig face recognition model. To train our model, 37 pigs with a total of 12,993 images were randomly selected from the collected pig face images, and 9 pigs with a total of 3431 images were set as a test set. 900 pairs of positive sample pairs and 900 pairs of negative pairs were obtained from the images in the test set. A series of experimental results show that our accuracy reached 95.28%, which was 2.61% higher than that of a human face recognition model. NAM was more effective in improving the performance of the pig face recognition model than the mainstream BAM (bottleneck attention module) and CBAM (convolutional block attention module). The research results can provide technological support for non-contact open-set individual recognition for intelligent farming processes.