Accuracy of the TurfTrax Racing Data System for determination of equine speed and position

Reasons for performing study: The speed and position data collected by TurfTrax Racing Data Limited during UK Thoroughbred racing have potential to benefit equine science and welfare. The size (the 2006 data set alone consists of 30,932 individual horse starts across 2667 races) and nature (speed and 2D position for each horse at 4 updates per second) of the data make it a unique resource for questions in equine safety, welfare, performance, and animal locomotion. Objective: To determine the accuracy of the TurfTrax tracking system in estimating the speed and position of horses during racing. Methods: Measurements from the TurfTrax wireless tracking system were compared with those of a survey‐grade global positioning system (GPS) receiver. Results: The TurfTrax system was found to give position measurements to within ± 11 and ± 64 cm in the fore‐aft and lateral directions, respectively, averaging ± 38 cm (interquartile range) and speed to within 0.15 m/s. Potential relevance: The data collected by the TurfTrax system are of sufficient accuracy to inform new diagnoses, training regimens and basic locomotor scientific studies. The position data can provide the precise distance, going, inclination, rate of turn and pack positioning through which each horse has raced. The speed profile can be used to examine the level of exertion, effect of training regimens and influence of racecourse features on performance. A first clinical application would be to analyse retrospectively these factors on occurrence of injury to compare with past training regimens, levels of exertion, and/or racecourse conditions.     

Fractals and search paths in mammals

The fractal index by Katz and George (1985) for thecharacterization of planar curves is applied to wolf search pathsrecorded by radio-telemetry. All the sets of paths studied showspatial patterns whose complexity is between a straight line anda true random walk. Females fractal dimensions show significantchanges throughout the year, depending on the state of their lifecycle (normal, breeding and wandering). There are alsodifferences between males and females, but not between adults andnon-adults. The results are discussed with regard to wolffood-search strategies.     

无线电追踪技术在地下啮齿动物研究中的应用

营地下生活的啮齿动物被誉为陆地生态系统的工程师,但其长期栖居于地下的生活习性难以被人们直接观察,给动物行为学和生态生物学等研究带来了挑战。随着科学技术的发展,无线电技术的小型化和轻量化为地下啮齿动物研究提供了可行性。国外学者已成功将无线电追踪技术应用于对地下啮齿动物洞道系统、活动节律、巢域面积变化以及社群制度等的研究。但国内将其应用于地下啮齿动物研究的较少。本文根据目前国内外研究现状,对无线追踪技术的研究内容和研究方法进行综述,以期为我国地下啮齿类研究提供参考。     

A model of the spring migration into the North Sea by Calanus finmarchicus overwintering off the Scottish continental shelf

A particle tracking model was used to investigate the annual spring invasion of the North Sea by Calanus finmarchicus copepodites which overwinter in deep water off the Scottish continental shelf. Flow fields generated by a hydrodynamic model (HDM) were used to simulate the advection of zero drag particles representing the copepods. Particles were released simultaneously from a regular lattice of start positions at a given depth ( D ₁), and ascended at a fixed rate ( dD/dt ) until they reached a final depth ( D ₂) in the surface layers. The proportion of particles reaching target areas in the northern North Sea was relatively insensitive to a moderate degree of variation (±20%) around chosen default values of the vertical migration parameters ( D ₁, D ₂ and dD/dt ), derived from field data. The inclusion of horizontal diffusion velocities and diel vertical migration in surface layers did not significantly affect the results. Sensitivity to wind direction was investigated by applying flow fields from HDM runs with different wind forcing scenarios. For the default vertical migration parameters, only north-westerly winds resulted in particles entering the North Sea from release locations north of the Iceland–Scotland Ridge, where dense aggregations of overwintering copepods were encountered during field surveys. The particle tracking model predicted that the major routes for the spring Calanus invasion into the North Sea were the East of Shetland Atlantic Inflow and the Norwegian Trench Atlantic Inflow, which agreed with seasonal trends observed in Continuous Plankton Recorder data. Overall, despite its relative simplicity, particle tracking was confirmed as a robust tool to explore the causal mechanisms behind the annual invasion of the North Sea by C. finmarchicus emerging from diapause in the deep waters off the Scottish continental shelf.     

A computer vision system to analyse the swimming behaviour of farmed fish in commercial aquaculture facilities: A case study using cage-held Atlantic salmon

Knowledge of fish behaviour plays an important role in aquaculture farm management. Video systems are the most common and cost-effective way of observing behaviours in commercial aquaculture operations. However long term observation is not feasible due to a limited ability to analyse footage manually. This paper describes preliminary findings obtained via computer vision software that was developed to automatically analyse fish movement and behaviours in aquaculture sea cages. Results show that the system is capable of detecting fish shapes in video recordings and from these shapes quantifying changes in swimming speed and direction continuously throughout the day. Also variations between days were detected and these may have been associated with the daily shift in the tidal cycle. The system has the potential to act as an alarm to farm operators, informing them about unusual fish behaviours on a continuous, real-time basis. It also has potential to assist in the evaluation of fish welfare.     

基于多尺度融合注意力机制的群猪检测方法

群猪检测是现代化猪场智慧管理的关键环节。针对群猪计数过程中,小目标或被遮挡的猪只个体易漏检的问题,该研究提出了基于多尺度融合注意力机制的群猪检测方法。首先基于YOLOv7模型构建了群猪目标检测网络YOLOpig,该网络设计了融合注意力机制的小目标尺度检测网络结构,并基于残差思想优化了最大池化卷积模块,实现了对被遮挡与小目标猪只个体的准确检测;其次结合GradCAM算法进行猪只检测信息的特征可视化,验证群猪检测试验特征提取的有效性。最后使用目标跟踪算法StrongSORT实现猪只个体的准确跟踪,为猪只的检测任务提供身份信息。研究以育肥阶段的长白猪为测试对象,基于不同视角采集的视频数据集进行测试,验证了YOLOpig网络结合StongSORT算法的准确性和实时性。试验结果表明,该研究提出的YOLOpig模型精确率、召回率及平均精度分别为90.4%、85.5%和92.4%,相较于基础YOLOv7模型平均精度提高了5.1个百分点,检测速度提升7.14%,比YOLOv5、YOLOv7tiny和YOLOv8n 3种模型的平均精度分别提高了12.1、16.8和5.7个百分点,该文模型可以实现群猪的有...     

拖拉机行走机构的研究现状

牵引附着性能是拖拉机最重要的性能之一,多年的研究表明,要提高牵引附着性能,进行保护性耕作,改善车辆在松软地面的通过性能,新的行走机构或装置就成为目前国内外研究的主要趋向。为此,介绍了国内外行走机构研究的现状,并指出了现存问题和今后发展的趋势。     

基于YOLO v8n-seg和改进Strongsort的多目标小鼠跟踪方法

多目标小鼠跟踪是小鼠行为分析的基本任务,是研究社交行为的重要方法。针对传统小鼠跟踪方法存在只能跟踪单只小鼠以及对多目标小鼠跟踪需要对小鼠进行标记从而影响小鼠行为等问题,提出了一种基于实例分割网络YOLO v8n-seg和改进Strongsort相结合的多目标小鼠无标记跟踪方法。使用RGB摄像头采集多目标小鼠的日常行为视频,标注小鼠身体部位分割数据集,对数据集进行增强后训练YOLO v8n-seg实例分割网络,经过测试,模型精确率为97.7%,召回率为98.2%,mAP50为99.2%,单幅图像检测时间为3.5ms,实现了对小鼠身体部位准确且快速地分割,可以满足Strongsort多目标跟踪算法的检测要求。针对Strongsort算法在多目标小鼠跟踪中存在的跟踪错误问题,对Strongsort做了两点改进：对匹配流程进行改进,将未匹配上目标的轨迹和未匹配上轨迹的目标按欧氏距离进行再次匹配;对卡尔曼滤波进行改进,将卡尔曼滤波中表示小鼠位置和运动状态的小鼠身体轮廓外接矩形框替换为以小鼠身体轮廓质心为中心、对角线为小鼠体宽的正方形框。经测试,改进后Strongsort算法的ID跳变数为14,MOTA为97.698%,IDF1为85.435%,MOTP为75.858%,与原Strongsort相比,ID跳变数减少88%,MOTA提升3.266个百分点,IDF1提升27.778个百分点,与Deepsort、ByteTrack和Ocsort相比,在MOTA和IDF1上均有显著提升,且ID跳变数大幅降低,结果表明改进Strongsort算法可以提高多目标无标记小鼠跟踪的稳定性和准确性,为小鼠社交行为分析提供了一种新的技术途径。     

基于ASTUKF的分布式农业车辆路面参数辨识方法

针对分布式驱动农业车辆在路面参数辨识过程中,因路面环境变化出现的状态模型误差和时变噪声,导致辨识结果发散的问题,提出了基于自适应强跟踪无迹卡尔曼滤波(Adaptive strong tracking unscented Kalman filter, ASTUKF)的辨识方法。与传统内燃机农业车辆相比,分布式驱动可以直接获取驱动轮的状态信息,结合含有峰值附着系数和极限滑转率的μ-s曲线模型,建立了无迹卡尔曼滤波(Unscented Kalman filter, UKF)辨识算法的状态方程和量测方程。同时,将强跟踪滤波(Strong tracking filter, STF)和自适应滤波(Adaptive filter, AF)引入辨识算法,用以提高对多变环境的识别精度和鲁棒性,并采用奇异值分解(Singular value decomposition, SVD)解决了迭代过程中出现的非正定矩阵的问题。仿真试验结果表明,在突变噪声环境工况下,ASTUKF辨识结果可以快速收敛至目标值附近,且不受突变噪声的影响,各驱动轮峰值附着系数估计结果的平均绝对误差(Mean absolute error...