At present,both the point source and the imaging polarization navigation devices only can output the angle information,which means that the velocity information of the carrier cannot be extracted from the polarization...At present,both the point source and the imaging polarization navigation devices only can output the angle information,which means that the velocity information of the carrier cannot be extracted from the polarization field pattern directly.Optical flow is an image-based method for calculating the velocity of pixel point movement in an image.However,for ordinary optical flow,the difference in pixel value as well as the calculation accuracy can be reduced in weak light.Polarization imaging technology has the ability to improve both the detection accuracy and the recognition probability of the target because it can acquire the extra polarization multi-dimensional information of target radiation or reflection.In this paper,combining the polarization imaging technique with the traditional optical flow algorithm,a polarization optical flow algorithm is proposed,and it is verified that the polarized optical flow algorithm has good adaptation in weak light and can improve the application range of polarization navigation sensors.This research lays the foundation for day and night all-weather polarization navigation applications in future.展开更多
Current deformation measurement techniques suffer from limited spatial resolution. In this work, a highly accurate and high-resolution Horn Schunck optical flow method is developed and then applied to measuring the st...Current deformation measurement techniques suffer from limited spatial resolution. In this work, a highly accurate and high-resolution Horn Schunck optical flow method is developed and then applied to measuring the static deformation of a birdlike flexible airfoil at a series of angles of attack at Reynolds number 100,000 in a low speed, low noise wind tunnel. To allow relatively large displacements, a nonlinear Horn-Schunck model and a coarse-to-fine warping process are adopted. To preserve optical flow discontinuities, a nonquadratic penalization function, a multi- cue driven bilateral filtering and a principle component analysis of local image patterns are used. First, the accuracy and convergence of this Horn-Schunck technique are verified on a benchmark. Then, the maximum displacement that can be reliably calculated by this technique is studied on synthetic images. Both studies are compared with the performance of a Lucas-Kanade optical flow method. Finally, the Horn-Schunck technique is used to estimate the 3-D deformation of the birdlike airfoil through a stereoscopic camera setup. The results are compared with those computed by Lucas-Kanade optical flow, image correlation and numerical simulation.展开更多
Sea ice velocity impacts the distribution of sea ice,and the flux of exported sea ice through the Fram Strait increases with increasing ice velocity.Therefore,improving the accuracy of estimates of the sea ice velocit...Sea ice velocity impacts the distribution of sea ice,and the flux of exported sea ice through the Fram Strait increases with increasing ice velocity.Therefore,improving the accuracy of estimates of the sea ice velocity is important.We introduce a pyramid algorithm into the Horn-Schunck optical flow(HS-OF)method(to develop the PHS-OF method).Before calculating the sea ice velocity,we generate multilayer pyramid images from an original brightness temperature image.Then,the sea ice velocity of the pyramid layer is calculated,and the ice velocity in the original image is calculated by layer iteration.Winter Arctic sea ice velocities from 2014 to 2016 are obtained and used to discuss the accuracy of the HS-OF method and PHS-OF(specifically the 2-layer PHS-OF(2 LPHS-OF)and 4-layer PHS-OF(4 LPHS-OF))methods.The results prove that the PHS-OF method indeed improves the accuracy of sea ice velocity estimates,and the 2 LPHS-OF scheme is more appropriate for estimating ice velocity.The error is smaller for the 2 LPHS-OF velocity estimates than values from the Ocean and Sea Ice Satellite Application Facility and the Copernicus Marine Environment Monitoring Service,and estimates of changes in velocity by the 2 LPHS-OF method are consistent with those from the National Snow and Ice Data Center.Sea ice undergoes two main motion patterns,i.e.,transpolar drift and the Beaufort Gyre.In addition,cyclonic and anticyclonic ice drift occurred during winter 2016.Variations in sea ice velocity are related to the open water area,sea ice retreat time and length of the open water season.展开更多
Particles occur in almost all processes in chemical and life sciences. The particle size and shape influence the process performance and product quality, and in turn they are influenced by the flow behavior of the par...Particles occur in almost all processes in chemical and life sciences. The particle size and shape influence the process performance and product quality, and in turn they are influenced by the flow behavior of the particles during production. Monitoring and controlling such characteristics in multiphase systems to obtain sufficient qualities will greatly facilitate the achievement of reproducible and defined distributions. So far, obtaining this information inline has been challenging, because existing instruments lack measurement precision, being unable to process overlapping signals from different particle phases in highly concentrated multiphase systems. However, recent advances in photo-optics made it possible to monitor such features(particle size distribution(PSD), aspect ratio and particle concentration) with advanced image analysis(IA) in real-time. New analysis workflows as well as single feature extractions from the images using multiple image analysis algorithms allowed the precise real-time measurements of size, shape and concentration of particle collectives even separated from each other in three phase systems. The performances, advantages and drawbacks with other non-photo-optical methods for assessing the particle size distribution are compared and discussed.展开更多
We have numerically and experimentally investigated the flow rate measurement of the pipeline based on the optical fiber.Employing the large eddy simulation(LES)model,we have quantitatively analyzed the pressure fluct...We have numerically and experimentally investigated the flow rate measurement of the pipeline based on the optical fiber.Employing the large eddy simulation(LES)model,we have quantitatively analyzed the pressure fluctuation of the pipe wall caused by the turbulent flow in the pipeline.The simulation results have shown that the standard deviation of pressure fluctuation was quadratic with the flow rate.We have verified the theoretical model by using a distributed optical fiber acoustic sensing(DAS)system in the flow rate range from 0.61 m/s to 2.42 m/s.The experimental results were consistent with the simulation results very well.Furthermore,to improve the measuring error at the low flow rate,we have employed the composite adaptive denoising algorithm to eliminate the background noise and system noise.The final results have shown that the minimum goodness of fit was improved from 0.962 to 0.997,and the variation of the quadratic coefficient significantly decreased by 93.25%.The measured flow rate difference was only 0.84%between different sensing points in repeated experiments.展开更多
针对多视觉任务中传输成本高、解码端计算压力大的问题,提出一种自适应可伸缩视频编码(adaptive scalable video coding,ASVC)传输框架,将视频分为语义层和背景层,分别传输语义和背景信息。此外,提出一种自适应压缩算法,构建了C4.5决策...针对多视觉任务中传输成本高、解码端计算压力大的问题,提出一种自适应可伸缩视频编码(adaptive scalable video coding,ASVC)传输框架,将视频分为语义层和背景层,分别传输语义和背景信息。此外,提出一种自适应压缩算法,构建了C4.5决策树模型分析网络环境对视频进行压缩的决策判定,并对帧序列进行光流分析,在保留变化显著的帧基础上引入插值机制保持图像的平滑性。仿真结果表明,ASVC方法在不同码率环境下表现更高的识别精准率,视频质量和传输效率的显著提升。展开更多
[目的]运用You Only Look Once version 5s(YOLOv5s)模型实现对分娩奶山羊躺卧姿态的自动识别,并结合Farneback光流算法对分娩奶山羊胸腹部起伏特征进行分析,从而为奶山羊分娩的精准化管理提供技术支撑。[方法]利用YOLOv5s模型对分娩奶...[目的]运用You Only Look Once version 5s(YOLOv5s)模型实现对分娩奶山羊躺卧姿态的自动识别,并结合Farneback光流算法对分娩奶山羊胸腹部起伏特征进行分析,从而为奶山羊分娩的精准化管理提供技术支撑。[方法]利用YOLOv5s模型对分娩奶山羊的躺卧与站立姿态进行分类识别,采用精确率(P)、召回率(R)及平均精确率(mAP)对模型分类结果进行评价。通过视频识别后,依据分娩时长将20只萨能奶山羊分为2组:A组为分娩时长<30 min,B组为分娩时长≥30 min。并基于Farneback光流算法提取分娩奶山羊胸腹部起伏参数(速度、高度、单次持续时间、次数),对比分析两组奶山羊胸腹部运动规律。[结果]①YOLOv5s模型对躺卧和站立姿态识别的P分别为98.4%和98.3%,假阳性率<2%,误判风险极低;R为95.3%和94.6%,漏检率<6%,监测覆盖性优异;mAP达96.3%和95.2%,综合性能稳定,鲁棒性强。②光流法分析表明,B组胸腹部起伏速度均值为5.358 px/s,显著(P<0.05)高于A组均值(2.461 px/s);B组胸腹部起伏高度均值为6.104 px,极显著(P<0.01)高于A组均值(2.280 px);B组单次起伏持续时间均值(4.687 s)与A组均值(4.272 s)差异不显著(P=0.35);B组胸腹部起伏次数(45.67次)极显著(P<0.01)高于A组(12.92次),且节律性降低,这表明分娩难度随分娩时长增加而升高。[结论]YOLOv5s模型与Farneback光流算法协同运作,实现了对奶山羊分娩姿态的精准识别以及胸腹部运动精准量化。该技术能够集成到牧场分娩预警系统中,实时识别奶山羊的异常分娩行为,降低母羊难产风险,为奶山羊的智能化管理提供技术支持。展开更多
针对ORB(oriented FAST and rotated BRIEF)算法中存在匹配精确率低的问题,提出了一种基于LK(Lucas-Kanade)光流改进的ORB图像匹配方法。首先对待处理的图像进行直方图均衡化,然后在Oriented FAST特征点检测的同时用LK光流对其进行跟踪...针对ORB(oriented FAST and rotated BRIEF)算法中存在匹配精确率低的问题,提出了一种基于LK(Lucas-Kanade)光流改进的ORB图像匹配方法。首先对待处理的图像进行直方图均衡化,然后在Oriented FAST特征点检测的同时用LK光流对其进行跟踪,并将跟踪的特征点进行Rotated BRIEF描述,最后在特征匹配筛选环节利用RANSAC(Random Sampling Consistency)算法进行误匹配的剔除。实验结果表明,改进算法在公开数据集中的平均匹配精度为90.9%,平均特征匹配及误匹配的剔除共耗时为18ms,与原始ORB算法相比,在时间基本一致的前提下,有效的提高了匹配的精度。展开更多
基金supported by the National Natural Science Foundation of China(Nos.51675076 and 51505062)the Science Fund for Creative Research Groups of NSFC(No.51621064)the Basic scientific research fees for Central Universities(Nos.DUT17GF109 and DUT16TD20)
文摘At present,both the point source and the imaging polarization navigation devices only can output the angle information,which means that the velocity information of the carrier cannot be extracted from the polarization field pattern directly.Optical flow is an image-based method for calculating the velocity of pixel point movement in an image.However,for ordinary optical flow,the difference in pixel value as well as the calculation accuracy can be reduced in weak light.Polarization imaging technology has the ability to improve both the detection accuracy and the recognition probability of the target because it can acquire the extra polarization multi-dimensional information of target radiation or reflection.In this paper,combining the polarization imaging technique with the traditional optical flow algorithm,a polarization optical flow algorithm is proposed,and it is verified that the polarized optical flow algorithm has good adaptation in weak light and can improve the application range of polarization navigation sensors.This research lays the foundation for day and night all-weather polarization navigation applications in future.
文摘Current deformation measurement techniques suffer from limited spatial resolution. In this work, a highly accurate and high-resolution Horn Schunck optical flow method is developed and then applied to measuring the static deformation of a birdlike flexible airfoil at a series of angles of attack at Reynolds number 100,000 in a low speed, low noise wind tunnel. To allow relatively large displacements, a nonlinear Horn-Schunck model and a coarse-to-fine warping process are adopted. To preserve optical flow discontinuities, a nonquadratic penalization function, a multi- cue driven bilateral filtering and a principle component analysis of local image patterns are used. First, the accuracy and convergence of this Horn-Schunck technique are verified on a benchmark. Then, the maximum displacement that can be reliably calculated by this technique is studied on synthetic images. Both studies are compared with the performance of a Lucas-Kanade optical flow method. Finally, the Horn-Schunck technique is used to estimate the 3-D deformation of the birdlike airfoil through a stereoscopic camera setup. The results are compared with those computed by Lucas-Kanade optical flow, image correlation and numerical simulation.
基金The National Key Research and Development Program of China under contract Nos 2018YFC1407200 and 2018YFC1407203the National Natural Science Foundation of China under contract No.41976212
文摘Sea ice velocity impacts the distribution of sea ice,and the flux of exported sea ice through the Fram Strait increases with increasing ice velocity.Therefore,improving the accuracy of estimates of the sea ice velocity is important.We introduce a pyramid algorithm into the Horn-Schunck optical flow(HS-OF)method(to develop the PHS-OF method).Before calculating the sea ice velocity,we generate multilayer pyramid images from an original brightness temperature image.Then,the sea ice velocity of the pyramid layer is calculated,and the ice velocity in the original image is calculated by layer iteration.Winter Arctic sea ice velocities from 2014 to 2016 are obtained and used to discuss the accuracy of the HS-OF method and PHS-OF(specifically the 2-layer PHS-OF(2 LPHS-OF)and 4-layer PHS-OF(4 LPHS-OF))methods.The results prove that the PHS-OF method indeed improves the accuracy of sea ice velocity estimates,and the 2 LPHS-OF scheme is more appropriate for estimating ice velocity.The error is smaller for the 2 LPHS-OF velocity estimates than values from the Ocean and Sea Ice Satellite Application Facility and the Copernicus Marine Environment Monitoring Service,and estimates of changes in velocity by the 2 LPHS-OF method are consistent with those from the National Snow and Ice Data Center.Sea ice undergoes two main motion patterns,i.e.,transpolar drift and the Beaufort Gyre.In addition,cyclonic and anticyclonic ice drift occurred during winter 2016.Variations in sea ice velocity are related to the open water area,sea ice retreat time and length of the open water season.
基金financially supported by the grants for the project "Smart Process Inspection" (funding code ZF4184501CR5) from the "Zentrales Innovationsprogramm Mittelstand" (ZIM)
文摘Particles occur in almost all processes in chemical and life sciences. The particle size and shape influence the process performance and product quality, and in turn they are influenced by the flow behavior of the particles during production. Monitoring and controlling such characteristics in multiphase systems to obtain sufficient qualities will greatly facilitate the achievement of reproducible and defined distributions. So far, obtaining this information inline has been challenging, because existing instruments lack measurement precision, being unable to process overlapping signals from different particle phases in highly concentrated multiphase systems. However, recent advances in photo-optics made it possible to monitor such features(particle size distribution(PSD), aspect ratio and particle concentration) with advanced image analysis(IA) in real-time. New analysis workflows as well as single feature extractions from the images using multiple image analysis algorithms allowed the precise real-time measurements of size, shape and concentration of particle collectives even separated from each other in three phase systems. The performances, advantages and drawbacks with other non-photo-optical methods for assessing the particle size distribution are compared and discussed.
基金supported in part by the National Natural Science Foundation of China(Grant No.U22A20206)the Key Research and Development Plan Project of Hubei Province,China(Grant No.2022BAA004)Zhejiang Provincial Market Supervision Bureau Young Eagle Plan Project,China(Grant No.CY2022228).
文摘We have numerically and experimentally investigated the flow rate measurement of the pipeline based on the optical fiber.Employing the large eddy simulation(LES)model,we have quantitatively analyzed the pressure fluctuation of the pipe wall caused by the turbulent flow in the pipeline.The simulation results have shown that the standard deviation of pressure fluctuation was quadratic with the flow rate.We have verified the theoretical model by using a distributed optical fiber acoustic sensing(DAS)system in the flow rate range from 0.61 m/s to 2.42 m/s.The experimental results were consistent with the simulation results very well.Furthermore,to improve the measuring error at the low flow rate,we have employed the composite adaptive denoising algorithm to eliminate the background noise and system noise.The final results have shown that the minimum goodness of fit was improved from 0.962 to 0.997,and the variation of the quadratic coefficient significantly decreased by 93.25%.The measured flow rate difference was only 0.84%between different sensing points in repeated experiments.
文摘针对多视觉任务中传输成本高、解码端计算压力大的问题,提出一种自适应可伸缩视频编码(adaptive scalable video coding,ASVC)传输框架,将视频分为语义层和背景层,分别传输语义和背景信息。此外,提出一种自适应压缩算法,构建了C4.5决策树模型分析网络环境对视频进行压缩的决策判定,并对帧序列进行光流分析,在保留变化显著的帧基础上引入插值机制保持图像的平滑性。仿真结果表明,ASVC方法在不同码率环境下表现更高的识别精准率,视频质量和传输效率的显著提升。
文摘[目的]运用You Only Look Once version 5s(YOLOv5s)模型实现对分娩奶山羊躺卧姿态的自动识别,并结合Farneback光流算法对分娩奶山羊胸腹部起伏特征进行分析,从而为奶山羊分娩的精准化管理提供技术支撑。[方法]利用YOLOv5s模型对分娩奶山羊的躺卧与站立姿态进行分类识别,采用精确率(P)、召回率(R)及平均精确率(mAP)对模型分类结果进行评价。通过视频识别后,依据分娩时长将20只萨能奶山羊分为2组:A组为分娩时长<30 min,B组为分娩时长≥30 min。并基于Farneback光流算法提取分娩奶山羊胸腹部起伏参数(速度、高度、单次持续时间、次数),对比分析两组奶山羊胸腹部运动规律。[结果]①YOLOv5s模型对躺卧和站立姿态识别的P分别为98.4%和98.3%,假阳性率<2%,误判风险极低;R为95.3%和94.6%,漏检率<6%,监测覆盖性优异;mAP达96.3%和95.2%,综合性能稳定,鲁棒性强。②光流法分析表明,B组胸腹部起伏速度均值为5.358 px/s,显著(P<0.05)高于A组均值(2.461 px/s);B组胸腹部起伏高度均值为6.104 px,极显著(P<0.01)高于A组均值(2.280 px);B组单次起伏持续时间均值(4.687 s)与A组均值(4.272 s)差异不显著(P=0.35);B组胸腹部起伏次数(45.67次)极显著(P<0.01)高于A组(12.92次),且节律性降低,这表明分娩难度随分娩时长增加而升高。[结论]YOLOv5s模型与Farneback光流算法协同运作,实现了对奶山羊分娩姿态的精准识别以及胸腹部运动精准量化。该技术能够集成到牧场分娩预警系统中,实时识别奶山羊的异常分娩行为,降低母羊难产风险,为奶山羊的智能化管理提供技术支持。
文摘针对ORB(oriented FAST and rotated BRIEF)算法中存在匹配精确率低的问题,提出了一种基于LK(Lucas-Kanade)光流改进的ORB图像匹配方法。首先对待处理的图像进行直方图均衡化,然后在Oriented FAST特征点检测的同时用LK光流对其进行跟踪,并将跟踪的特征点进行Rotated BRIEF描述,最后在特征匹配筛选环节利用RANSAC(Random Sampling Consistency)算法进行误匹配的剔除。实验结果表明,改进算法在公开数据集中的平均匹配精度为90.9%,平均特征匹配及误匹配的剔除共耗时为18ms,与原始ORB算法相比,在时间基本一致的前提下,有效的提高了匹配的精度。
