Interactive Dairy Goat Image Segmentation for Precision Livestock Farming.

Semantic segmentation and instance segmentation based on deep learning play a significant role in intelligent dairy goat farming. However, these algorithms require a large amount of pixel-level dairy goat image annotations for model training. At present, users mainly use Labelme for pixel-level annotation of images, which makes it quite inefficient and time-consuming to obtain a high-quality annotation result. To reduce the annotation workload of dairy goat images, we propose a novel interactive segmentation model called UA-MHFF-DeepLabv3+, which employs layer-by-layer multi-head feature fusion (MHFF) and upsampling attention (UA) to improve the segmentation accuracy of the DeepLabv3+ on object boundaries and small objects. Experimental results show that our proposed model achieved state-of-the-art segmentation accuracy on the validation set of DGImgs compared with four previous state-of-the-art interactive segmentation models, and obtained 1.87 and 4.11 on mNoC@85 and mNoC@90, which are significantly lower than the best performance of the previous models of 3 and 5. Furthermore, to promote the implementation of our proposed algorithm, we design and develop a dairy goat image-annotation system named DGAnnotation for pixel-level annotation of dairy goat images. After the test, we found that it just takes 7.12 s to annotate a dairy goat instance with our developed DGAnnotation, which is five times faster than Labelme.

A Method for Obtaining 3D Point Cloud Data by Combining 2D Image Segmentation and Depth Information of Pigs.

This paper proposes a method for automatic pig detection and segmentation using RGB-D data for precision livestock farming. The proposed method combines the enhanced YOLOv5s model with the Res2Net bottleneck structure, resulting in improved fine-grained feature extraction and ultimately enhancing the precision of pig detection and segmentation in 2D images. Additionally, the method facilitates the acquisition of 3D point cloud data of pigs in a simpler and more efficient way by using the pig mask obtained in 2D detection and segmentation and combining it with depth information. To evaluate the effectiveness of the proposed method, two datasets were constructed. The first dataset consists of 5400 images captured in various pig pens under diverse lighting conditions, while the second dataset was obtained from the UK. The experimental results demonstrated that the improved YOLOv5s_Res2Net achieved a mAP@0.5:0.95 of 89.6% and 84.8% for both pig detection and segmentation tasks on our dataset, while achieving a mAP@0.5:0.95 of 93.4% and 89.4% on the Edinburgh pig behaviour dataset. This approach provides valuable insights for improving pig management, conducting welfare assessments, and estimating weight accurately.

Cattle Target Segmentation Method in Multi-Scenes Using Improved DeepLabV3+ Method.

Obtaining animal regions and the relative position relationship of animals in the scene is conducive to further studying animal habits, which is of great significance for smart animal farming. However, the complex breeding environment still makes detection difficult. To address the problems of poor target segmentation effects and the weak generalization ability of existing semantic segmentation models in complex scenes, a semantic segmentation model based on an improved DeepLabV3+ network (Imp-DeepLabV3+) was proposed. Firstly, the backbone network of the DeepLabV3+ model was replaced by MobileNetV2 to enhance the feature extraction capability of the model. Then, the layer-by-layer feature fusion method was adopted in the Decoder stage to integrate high-level semantic feature information with low-level high-resolution feature information at multi-scale to achieve more precise up-sampling operation. Finally, the SENet module was further introduced into the network to enhance information interaction after feature fusion and improve the segmentation precision of the model under complex datasets. The experimental results demonstrate that the Imp-DeepLabV3+ model achieved a high pixel accuracy (PA) of 99.4%, a mean pixel accuracy (MPA) of 98.1%, and a mean intersection over union (MIoU) of 96.8%. Compared to the original DeepLabV3+ model, the segmentation performance of the improved model significantly improved. Moreover, the overall segmentation performance of the Imp-DeepLabV3+ model surpassed that of other commonly used semantic segmentation models, such as Fully Convolutional Networks (FCNs), Lite Reduced Atrous Spatial Pyramid Pooling (LR-ASPP), and U-Net. Therefore, this study can be applied to the field of scene segmentation and is conducive to further analyzing individual information and promoting the development of intelligent animal farming.

Frame vibration states identification for corn harvester based on joint improved empirical mode decomposition - Support vector machine method.

The frame of corn harvester is prone to vibration bending and torsional deformation due to the vibration caused by field road bumps and fluctuations. It poses a serious challenge to the reliability of machinery. Therefore it is critical to explore the vibration mechanism, and to identify the vibration states under different working conditions. To address the above problem, a vibration state identification method is proposed in this paper. An improved empirical mode decomposition (EMD) algorithm was used to decrease noise for signals of high noise and non-stationary vibration in the field. The support vector machine (SVM) model was used for identification of frame vibration states under different working conditions. The results showed that: (1) an improved EMD algorithm could effectively reduce noise interference and restore the effective information of the original signal. (2) based on improved EMD - SVM method identify the vibration states of the frame with the accuracy of 99.21%. (3) The corn ears in grain tank were not sensitive to low order vibration, but had an absorption effect on high order vibration. The proposed method has the potential to be applied for accurately identifying vibration state and improving frame safety.

LettuceTrack: Detection and tracking of lettuce for robotic precision spray in agriculture.

The precision spray of liquid fertilizer and pesticide to plants is an important task for agricultural robots in precision agriculture. By reducing the amount of chemicals being sprayed, it brings in a more economic and eco-friendly solution compared to conventional non-discriminated spray. The prerequisite of precision spray is to detect and track each plant. Conventional detection or segmentation methods detect all plants in the image captured under the robotic platform, without knowing the ID of the plant. To spray pesticides to each plant exactly once, tracking of every plant is needed in addition to detection. In this paper, we present LettuceTrack, a novel Multiple Object Tracking (MOT) method to simultaneously detect and track lettuces. When the ID of each plant is obtained from the tracking method, the robot knows whether a plant has been sprayed before therefore it will only spray the plant that has not been sprayed. The proposed method adopts YOLO-V5 for detection of the lettuces, and a novel plant feature extraction and data association algorithms are introduced to effectively track all plants. The proposed method can recover the ID of a plant even if the plant moves out of the field of view of camera before, for which existing Multiple Object Tracking (MOT) methods usually fail and assign a new plant ID. Experiments are conducted to show the effectiveness of the proposed method, and a comparison with four state-of-the-art Multiple Object Tracking (MOT) methods is shown to prove the superior performance of the proposed method in the lettuce tracking application and its limitations. Though the proposed method is tested with lettuce, it can be potentially applied to other vegetables such as broccoli or sugar beat.

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.

Multi-leveled insights into the response of the eelgrass Zostera marina L to Cu than Cd exposure.

Seagrass beds are recognized as critical and among the most vulnerable habitats on the planet; seagrass colonize the coastal waters where heavy metal pollution is a serious problem. In this study, the toxic effects of copper and cadmium in the eelgrass Zostera marina L. were observed at the individual, subcellular, physiologically biochemical, and molecular levels. Both Cu and Cd stress significantly inhibited the growth and the maximal quantum yield of photosystem II (Fv/Fm); and high temperature increased the degree of heavy metal damage, while low temperatures inhibited damage. The half-effect concentration (EC50) of eelgrass was 28.9 µM for Cu and 2246.8 µM for Cd, indicating Cu was much more toxic to eelgrass than Cd. The effect of Cu and Cd on photosynthesis was synergistic. After 14 days of enrichment, the concentration of Cu in leaves and roots of Z. marina was 48 and 37 times higher than that in leaf sheath, and 14 and 11 times higher than that in rhizome; and the order of Cd concentration in the organs was root > leaf > rhizome > sheath. Heavy metal uptake mainly occurred in the organelles, and Cd enrichment also occurred to a certain extent in the cytoplasm. Transcriptome results showed that a number of photosynthesis-related KEGG enrichment pathways and GO terms were significantly down-regulated under Cd stress, suggesting that the photosynthetic system of eelgrass was severely damaged at the transcriptome level, which was consistent with the significant inhibition of Fv/Fm and leaf yellowing. Under Cu stress, the genes related to glutathione metabolic pathway were significantly up-regulated, together with the increased autioxidant enzyme activity of GSH-PX. In addition, the results of recovery experiment indicated that the damage caused by short-term Cd and Cu stress under EC50 was reversible. These results provide heavy metal toxic effects at multiple levels and information relating to the heavy metal resistance strategies evolved by Z. marina to absorb and isolate heavy metals, and highlight the phytoremediation potential of this species especially for Cd.

Deep Learning Based Automatic Grape Downy Mildew Detection.

Grape downy mildew (GDM) disease is a common plant leaf disease, and it causes serious damage to grape production, reducing yield and fruit quality. Traditional manual disease detection relies on farm experts and is often time-consuming. Computer vision technologies and artificial intelligence could provide automatic disease detection for real-time controlling the spread of disease on the grapevine in precision viticulture. To achieve the best trade-off between GDM detection accuracy and speed under natural environments, a deep learning based approach named YOLOv5-CA is proposed in this study. Here coordinate attention (CA) mechanism is integrated into YOLOv5, which highlights the downy mildew disease-related visual features to enhance the detection performance. A challenging GDM dataset was acquired in a vineyard under a nature scene (consisting of different illuminations, shadows, and backgrounds) to test the proposed approach. Experimental results show that the proposed YOLOv5-CA achieved a detection precision of 85.59%, a recall of 83.70%, and a mAP@0.5 of 89.55%, which is superior to the popular methods, including Faster R-CNN, YOLOv3, and YOLOv5. Furthermore, our proposed approach with inference occurring at 58.82 frames per second, could be deployed for the real-time disease control requirement. In addition, the proposed YOLOv5-CA based approach could effectively capture leaf disease related visual features resulting in higher GDE detection accuracy. Overall, this study provides a favorable deep learning based approach for the rapid and accurate diagnosis of grape leaf diseases in the field of automatic disease detection.

Do adult eelgrass shoots rule seedling fate in a large seagrass meadow in a eutrophic bay in northern China?

We conducted field sampling over 19 months to investigate eelgrass population reproduction status and ecological interactions in a large seagrass meadow in a eutrophic bay in northern China. The results showed asexual growth played an important role in the maintenance of existing meadows, and sexual reproduction played a critical role in the colonization of new areas. We conclude that adult eelgrass shoots do rule the fate of seedlings in the large seagrass meadow. Additionally, nutrient resources (N and P) at this location were found to meet eelgrass growth demand. The N/P ratios of seawater and seagrass indicated N limitation relative to P in the eutrophic bay based on the seagrass Redfield ratio (25-30). Nutrient uptake by seagrass might be an important factor in reducing the probability of a red tide in the study area. The results of this study provide fundamental information for eelgrass restoration and conservation.

One-Shot Learning with Pseudo-Labeling for Cattle Video Segmentation in Smart Livestock Farming.

Computer vision-based technologies play a key role in precision livestock farming, and video-based analysis approaches have been advocated as useful tools for automatic animal monitoring, behavior analysis, and efficient welfare measurement management. Accurately and efficiently segmenting animals' contours from their backgrounds is a prerequisite for vision-based technologies. Deep learning-based segmentation methods have shown good performance through training models on a large amount of pixel-labeled images. However, it is challenging and time-consuming to label animal images due to their irregular contours and changing postures. In order to reduce the reliance on the number of labeled images, one-shot learning with a pseudo-labeling approach is proposed using only one labeled image frame to segment animals in videos. The proposed approach is mainly comprised of an Xception-based Fully Convolutional Neural Network (Xception-FCN) module and a pseudo-labeling (PL) module. Xception-FCN utilizes depth-wise separable convolutions to learn different-level visual features and localize dense prediction based on the one single labeled frame. Then, PL leverages the segmentation results of the Xception-FCN model to fine-tune the model, leading to performance boosts in cattle video segmentation. Systematic experiments were conducted on a challenging feedlot cattle video dataset acquired by the authors, and the proposed approach achieved a mean intersection-over-union score of 88.7% and a contour accuracy of 80.8%, outperforming state-of-the-art methods (OSVOS and OSMN). Our proposed one-shot learning approach could serve as an enabling component for livestock farming-related segmentation and detection applications.

Maize disease detection based on spectral recovery from RGB images.

Maize is susceptible to infect pest disease, and early disease detection is key to preventing the reduction of maize yields. The raw data used for plant disease detection are commonly RGB images and hyperspectral images (HSI). RGB images can be acquired rapidly and low-costly, but the detection accuracy is not satisfactory. On the contrary, using HSIs tends to obtain higher detection accuracy, but HSIs are difficult and high-cost to obtain in field. To overcome this contradiction, we have proposed the maize spectral recovery disease detection framework which includes two parts: the maize spectral recovery network based on the advanced hyperspectral recovery convolutional neural network (HSCNN+) and the maize disease detection network based on the convolutional neural network (CNN). Taking raw RGB data as input of the framework, the output reconstructed HSIs are used as input of disease detection network to achieve disease detection task. As a result, the detection accuracy obtained by using the low-cost raw RGB data almost as same as that obtained by using HSIs directly. The HSCNN+ is found to be fit to our spectral recovery model and the reconstruction fidelity was satisfactory. Experimental results demonstrate that the reconstructed HSIs efficiently improve detection accuracy compared with raw RGB image in tested scenarios, especially in complex environment scenario, for which the detection accuracy increases by 6.14%. The proposed framework has the advantages of fast, low cost and high detection precision. Moreover, the framework offers the possibility of real-time and precise field disease detection and can be applied in agricultural robots.

Intelligent Perception-Based Cattle Lameness Detection and Behaviour Recognition: A Review.

The growing world population has increased the demand for animal-sourced protein. However, animal farming productivity is faced with challenges from traditional farming practices, socioeconomic status, and climate change. In recent years, smart sensors, big data, and deep learning have been applied to animal welfare measurement and livestock farming applications, including behaviour recognition and health monitoring. In order to facilitate research in this area, this review summarises and analyses some main techniques used in smart livestock farming, focusing on those related to cattle lameness detection and behaviour recognition. In this study, more than 100 relevant papers on cattle lameness detection and behaviour recognition have been evaluated and discussed. Based on a review and a comparison of recent technologies and methods, we anticipate that intelligent perception for cattle behaviour and welfare monitoring will develop towards standardisation, a larger scale, and intelligence, combined with Internet of things (IoT) and deep learning technologies. In addition, the key challenges and opportunities of future research are also highlighted and discussed.

The super typhoon Lekima (2019) resulted in massive losses in large seagrass (Zostera japonica) meadows, soil organic carbon and nitrogen pools in the intertidal Yellow River Delta, China.

Seagrass meadows are key ecosystems, and they are among the most threatened habitats on the planet. Increased numbers of extreme climate events, such as hurricanes and marine heatwaves have caused severe damage to global seagrass meadows. The largest Zostera japonica meadows in China are located in the Yellow River Delta. It had a distribution area of 1031.8 ha prior to August 2019 when the Yellow River Delta was severely impacted by the passage of typhoon Lekima. In this study, we compared field data collected before and after the typhoon to determine its impact on seagrass beds in the Yellow River Delta. The super typhoon caused dramatic changes in Z. japonica in the Yellow River Delta, resulting in a greater than 100-fold decrease in distribution area, a greater than 35% loss of soil organic carbon, and a greater than 65% loss of soil total nitrogen in the top 35 cm sediments. Owing to the lack of seeds and overwintering shoots, as well as the small remaining distribution area, recovery was impossible, even though environmental factors were still suitable for species growth. Thus, restoration efforts are required for seagrass meadow recovery. Additionally, the long-term monitoring of this meadow will provide new information on the ecosystem's status and will be useful for future protection.

Can the Non-native Salt Marsh Halophyte Spartina alterniflora Threaten Native Seagrass (Zostera japonica) Habitats? A Case Study in the Yellow River Delta, China.

Seagrass meadows are critical ecosystems, and they are among the most threatened habitats on the planet. As an anthropogenic biotic invader, Spartina alterniflora Loisel. competes with native plants, threatens native ecosystems and coastal aquaculture, and may cause local biodiversity to decline. The distribution area of the exotic species S. alterniflora in the Yellow River Delta had been expanding to ca.4,000 ha from 1990 to 2018. In this study, we reported, for the first time, the competitive effects of the exotic plant (S. alterniflora) on seagrass (Zostera japonica Asch. & Graebn.) by field investigation and a transplant experiment in the Yellow River Delta. Within the first 3 months of the field experiment, S. alterniflora had pushed forward 14 m into the Z. japonica distribution region. In the study region, the area of S. alterniflora in 2019 increased by 516 times compared with its initial area in 2015. Inhibition of Z. japonica growth increased with the invasion of S. alterniflora. Z. japonica had been degrading significantly under the pressure of S. alterniflora invasion. S. alterniflora propagates sexually via seeds for long distance invasion and asexually by tillers and rhizomes for short distance invasion. Our results describe the invasion pattern of S. alterniflora and can be used to develop strategies for prevention and control of S. alterniflora invasion.

Diversity, distribution and conservation of seagrass in coastal waters of the Liaodong Peninsula, North Yellow Sea, northern China: Implications for seagrass conservation.

Seagrass beds are highly productive coastal ecosystems that are widely distributed along temperate and tropical coastlines globally. Although seagrass distribution and diversity have been widely reported on a global scale, there have been few reports on seagrass distribution and diversity in northern China, especially for coastal waters of the Liaodong Peninsula in the North Yellow Sea. In the present study, we investigated the distribution and diversity of seagrass in coastal waters of the Liaodong Peninsula in the North Yellow Sea, northern China. Field surveys of seagrass wrack were conducted along shorelines, to identify whether seagrass beds occurred in nearby waters, and sonar methods were then used to collect data relating to seagrass bed extent. Also, we analyzed the major threats facing seagrass beds. The results of the study revealed that four species (Zostera marina L., Z. japonica Aschers. & Graebn., Z. caespitosa M., and Phyllospadix iwatensis M.) were found in study area, covering a total area of 1253.47 ha. Seagrass bed area significantly decreased with increasing water depth, and most seagrass was recorded at depths of 2-5 m. Due to the steep slope of the seabed, seagrass beds exhibited a zonal distribution in most of the study areas. In addition, the amount of seagrass wrack along shorelines could be used to infer the size and distance of seagrass beds. Human activities, such as clam harvesting, land reclamation, coastal aquaculture pose a threat to the seagrass beds. This study provides new information to fill knowledge gaps regarding seagrass distribution in northern China and it provides a baseline for further monitoring of these seagrass beds.

New insights into physiological effects of anoxia under darkness on the iconic seagrass Zostera marina based on a combined analysis of transcriptomics and metabolomics.

Coastal hypoxia/anoxia is a major emerging threat to global coastal ecosystems. Macroalgae blooms of tens of kilometers are often observed in open waters. These blooms not only cause a lack of oxygen, but also benthic light limitation. We explored the physiological responses of Zostera marina L. to anoxia under darkness. After exposing Z. marina to anoxia under darkness for 72 h, we measured the elongation of leaves and the decrease in maximal quantum yield of photosystem II (Fv/Fm), and investigated the transcriptomic and metabolomic responses to anoxic stress based on RNA-sequencing and liquid chromatography-mass spectrometry (LC-MS) technology. The results showed that anoxic stress significantly reduced the leaf Fv/Fm, and had a significant negative effect on the photosynthesis and growth of Z. marina. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of up-regulated differentially expressed genes (DEGs) showed that glycolysis was the most significant enrichment pathway (p < 0.001), and most of the important products in glycolysis were significantly up-regulated. This indicated that the glycolysis process of anaerobic respiration is promoted under anoxia. The metabolite results also showed that glyceraldehyde 3-phosphate in the glycolysis pathway was significantly up-regulated. Moreover, three genes encoding sucrose synthase (gene-ZOSMA_310G00150, gene-ZOSMA_81G00980, and gene-ZOSMA_8G00730) and one gene encoding alpha-amylase (gene-ZOSMA_95G00270) were significantly up-regulated, providing the sugar basis for the subsequent increase in glycolysis. Furthermore, gene-encoding oxoglutarate dehydrogenase, the rate-limiting step of the tricarboxylic acid (TCA) cycle, was significantly down-regulated, indicating that this cycle was inhibited under anoxia. Metabolomic results showed that L-tryptophan, L-phenylalanine, and DL-leucine were significantly up-regulated. Only significantly decreased glutamate and non-significantly decreased glutamine, substances consumed in alanine and γ-aminobutyric acid (GABA) shunt mechanisms, were detected in the leaves, while GABA and alanine were not detected. The results of this study show that anoxic stress induces a programmed transcriptomic and metabolomic response in seagrass, most likely reflecting a complex strategy of acclimation and adaptation in seagrass to resist anoxic stress.