Reintroduction is an important strategy to restore or re-establish wild populations of endangered species.Pre-release training is a necessary step to ensure postreintroduction survival.However,studies reported contrad...Reintroduction is an important strategy to restore or re-establish wild populations of endangered species.Pre-release training is a necessary step to ensure postreintroduction survival.However,studies reported contradicting outcomes after pre-release training of juveniles and adults.This study used farmed and feral American mink(Neovison vison)to analyze the influence of captive breeding on the morphology,structure and efficiency of the two major hindlimb levers,the femur and tibia pivoted by hip and knee joints that are essential for locomotion.Results showed that captive breeding did not alter the sexual dimorphism of the two levers that are related to survival in the wild.Captive-bred mink showed slightly altered morphology of the femur and fundamental structure of the hindlimb levers that improved efficiency,but this resulted in reduction of performance related to foraging in both terrestrial and aquatic environments,especially for females.These findings suggest that reintroduction of mustelid as exampled by the mink here should focus on juveniles because the skeletal alterations associated with captive rearing were recorded only among adults and are irreversible in adulthood.In contrast,captive-reared juveniles showed no skeletal alterations and would be expected to recovery from any atrophy of the muscular system caused by captive rearing for shorter durations.Our results support the application of pre-release training of juveniles in enriched environments as a method for alleviating structural alteration of appendages and enhancing locomotion to increase survival probability in complex habitats.展开更多
We propose the superconducting van der Waals material 4Hb-TaS_(2)to realize the Z_(2)topological order and interpret the recent discovery of the spontaneous vortex generation in 4Hb-TaS_(2)as the vison-vortex nucleati...We propose the superconducting van der Waals material 4Hb-TaS_(2)to realize the Z_(2)topological order and interpret the recent discovery of the spontaneous vortex generation in 4Hb-TaS_(2)as the vison-vortex nucleation.For the alternating stacking of metallic/superconducting and Mott insulating layers in 4Hb-TaS_(2),we expect the local moments in the Mott insulating 1T-TaS_(2)layer to form the Z_(2)topological order.The spontaneous vortex generation in 4Hb-TaS_(2)is interpreted from the transition or nucleation between the superconducting vortex and the Z_(2)vison in different phase regimes.Differing from the single vison-vortex nucleation in the original Senthil-Fisher’s cuprate proposal,we consider such nucleation process between the superconducting vortex lattice and the vison crystal.We further propose experiments to distinguish this proposal with the Z_(2)topological order from the chiral spin liquid scenarios.展开更多
为了实现光伏硅片的缺陷类型的准确检测,提出一种基于改进YOLOv5(you only look once vison 5)的光伏硅片缺陷检测方法。首先对工厂生产的光伏硅片图像进行预处理,主要是给图片打上标签,制作成数据集。接着将其在YOLOv5s、YOLOv5m、YOLO...为了实现光伏硅片的缺陷类型的准确检测,提出一种基于改进YOLOv5(you only look once vison 5)的光伏硅片缺陷检测方法。首先对工厂生产的光伏硅片图像进行预处理,主要是给图片打上标签,制作成数据集。接着将其在YOLOv5s、YOLOv5m、YOLOv5n三种YOLOv5神经网络进行中训练,在测试集中进行验证,比较各神经网络的效果。最后针对光伏硅片斑点型缺陷效果不佳的问题,在神经网络中添加小目标检测层和SE注意力模块实现对光伏硅片斑点类型缺陷的准确检测,由实验结果可得,查准率高达97.4%,查全率高达97.8%,平均每张图片检测时间仅为21 ms。展开更多
为了解决落叶等杂物飘落到雨量筒网罩得不到及时清理导致雨量观测数据缺失或异常,影响降水统计和评估准确性的问题,基于YOLOv5(you only look once vison 5)目标检测算法开发并建立了一个雨量筒异物检测识别报警系统,首次将图像识别技...为了解决落叶等杂物飘落到雨量筒网罩得不到及时清理导致雨量观测数据缺失或异常,影响降水统计和评估准确性的问题,基于YOLOv5(you only look once vison 5)目标检测算法开发并建立了一个雨量筒异物检测识别报警系统,首次将图像识别技术应用在气象装备的维护维修领域。通过该系统对山东省区域自动气象站远程控制和实景监控系统实时抓拍的雨量筒照片进行图像识别,实现了对雨量筒中的异物检测。经实验统计,系统目前已检测识别1万余张雨量筒监控图片,对是否存在异物的识别准确率超过90%。此外,该系统还能够根据判断结果给出提醒,使工作人员可以及时准确地确定问题站点并清理维护。整体而言,该系统在一定程度上能够提高雨量观测的准确度。展开更多
For some important object recognition applications such as intelligent robots and unmanned driving, images are collected on a consecutive basis and associated among themselves, besides, the scenes have steady prior fe...For some important object recognition applications such as intelligent robots and unmanned driving, images are collected on a consecutive basis and associated among themselves, besides, the scenes have steady prior features. Yet existing technologies do not take full advantage of this information. In order to take object recognition further than existing algorithms in the above application, an object recognition method that fuses temporal sequence with scene priori information is proposed. This method first employs YOLOv3 as the basic algorithm to recognize objects in single-frame images, then the DeepSort algorithm to establish association among potential objects recognized in images of different moments, and finally the confidence fusion method and temporal boundary processing method designed herein to fuse, at the decision level, temporal sequence information with scene priori information. Experiments using public datasets and self-built industrial scene datasets show that due to the expansion of information sources, the quality of single-frame images has less impact on the recognition results, whereby the object recognition is greatly improved. It is presented herein as a widely applicable framework for the fusion of information under multiple classes. All the object recognition algorithms that output object class, location information and recognition confidence at the same time can be integrated into this information fusion framework to improve performance.展开更多
基金funded by China State Forestry and Grassland Administration Project for Rescue and Captive Breeding of Endangered and Rare Wildlife(2018).
文摘Reintroduction is an important strategy to restore or re-establish wild populations of endangered species.Pre-release training is a necessary step to ensure postreintroduction survival.However,studies reported contradicting outcomes after pre-release training of juveniles and adults.This study used farmed and feral American mink(Neovison vison)to analyze the influence of captive breeding on the morphology,structure and efficiency of the two major hindlimb levers,the femur and tibia pivoted by hip and knee joints that are essential for locomotion.Results showed that captive breeding did not alter the sexual dimorphism of the two levers that are related to survival in the wild.Captive-bred mink showed slightly altered morphology of the femur and fundamental structure of the hindlimb levers that improved efficiency,but this resulted in reduction of performance related to foraging in both terrestrial and aquatic environments,especially for females.These findings suggest that reintroduction of mustelid as exampled by the mink here should focus on juveniles because the skeletal alterations associated with captive rearing were recorded only among adults and are irreversible in adulthood.In contrast,captive-reared juveniles showed no skeletal alterations and would be expected to recovery from any atrophy of the muscular system caused by captive rearing for shorter durations.Our results support the application of pre-release training of juveniles in enriched environments as a method for alleviating structural alteration of appendages and enhancing locomotion to increase survival probability in complex habitats.
基金supported by the Ministry of Science and Technology of China with Grants No.2021YFA1400300the National Science Foundation of China with Grant No.92065203by the Fundamental Research Funds for the Central Universities,Peking University.
文摘We propose the superconducting van der Waals material 4Hb-TaS_(2)to realize the Z_(2)topological order and interpret the recent discovery of the spontaneous vortex generation in 4Hb-TaS_(2)as the vison-vortex nucleation.For the alternating stacking of metallic/superconducting and Mott insulating layers in 4Hb-TaS_(2),we expect the local moments in the Mott insulating 1T-TaS_(2)layer to form the Z_(2)topological order.The spontaneous vortex generation in 4Hb-TaS_(2)is interpreted from the transition or nucleation between the superconducting vortex and the Z_(2)vison in different phase regimes.Differing from the single vison-vortex nucleation in the original Senthil-Fisher’s cuprate proposal,we consider such nucleation process between the superconducting vortex lattice and the vison crystal.We further propose experiments to distinguish this proposal with the Z_(2)topological order from the chiral spin liquid scenarios.
文摘为了实现光伏硅片的缺陷类型的准确检测,提出一种基于改进YOLOv5(you only look once vison 5)的光伏硅片缺陷检测方法。首先对工厂生产的光伏硅片图像进行预处理,主要是给图片打上标签,制作成数据集。接着将其在YOLOv5s、YOLOv5m、YOLOv5n三种YOLOv5神经网络进行中训练,在测试集中进行验证,比较各神经网络的效果。最后针对光伏硅片斑点型缺陷效果不佳的问题,在神经网络中添加小目标检测层和SE注意力模块实现对光伏硅片斑点类型缺陷的准确检测,由实验结果可得,查准率高达97.4%,查全率高达97.8%,平均每张图片检测时间仅为21 ms。
文摘为了解决落叶等杂物飘落到雨量筒网罩得不到及时清理导致雨量观测数据缺失或异常,影响降水统计和评估准确性的问题,基于YOLOv5(you only look once vison 5)目标检测算法开发并建立了一个雨量筒异物检测识别报警系统,首次将图像识别技术应用在气象装备的维护维修领域。通过该系统对山东省区域自动气象站远程控制和实景监控系统实时抓拍的雨量筒照片进行图像识别,实现了对雨量筒中的异物检测。经实验统计,系统目前已检测识别1万余张雨量筒监控图片,对是否存在异物的识别准确率超过90%。此外,该系统还能够根据判断结果给出提醒,使工作人员可以及时准确地确定问题站点并清理维护。整体而言,该系统在一定程度上能够提高雨量观测的准确度。
文摘For some important object recognition applications such as intelligent robots and unmanned driving, images are collected on a consecutive basis and associated among themselves, besides, the scenes have steady prior features. Yet existing technologies do not take full advantage of this information. In order to take object recognition further than existing algorithms in the above application, an object recognition method that fuses temporal sequence with scene priori information is proposed. This method first employs YOLOv3 as the basic algorithm to recognize objects in single-frame images, then the DeepSort algorithm to establish association among potential objects recognized in images of different moments, and finally the confidence fusion method and temporal boundary processing method designed herein to fuse, at the decision level, temporal sequence information with scene priori information. Experiments using public datasets and self-built industrial scene datasets show that due to the expansion of information sources, the quality of single-frame images has less impact on the recognition results, whereby the object recognition is greatly improved. It is presented herein as a widely applicable framework for the fusion of information under multiple classes. All the object recognition algorithms that output object class, location information and recognition confidence at the same time can be integrated into this information fusion framework to improve performance.
