In light of degradation of particle filtering and robust weakness in the utilization of single feature tracking,this paper presents a kernel particle filtering tracking method based on multi-feature integration.In thi...In light of degradation of particle filtering and robust weakness in the utilization of single feature tracking,this paper presents a kernel particle filtering tracking method based on multi-feature integration.In this paper,a new weight upgrading method is given out during kernel particle filtering at first,and then robust tracking is realized by integrating color and texture features under the framework of kernel particle filtering.Space histogram and integral histogram is adopted to calculate color and texture features respectively.These two calculation methods effectively overcome their own defectiveness,and meanwhile,improve the real timing for particle filtering.This algorithm has also improved sampling effectiveness,resolved redundant calculation for particle filtering and degradation of particles.Finally,the experiment for target tracking is realized by using the method under complicated background and shelter.Experiment results show that the method can reliably and accurately track target and deal with target sheltering situation properly.展开更多
A novel particle filter bandwidth adaption for kernel particle filter(BAKPF)is proposed.Selection of the kernel bandwidth is a critical issue in kernel density estimation(KDE).The plug-in method is adopted to get the ...A novel particle filter bandwidth adaption for kernel particle filter(BAKPF)is proposed.Selection of the kernel bandwidth is a critical issue in kernel density estimation(KDE).The plug-in method is adopted to get the global fixed bandwidth by optimizing the asymptotic mean integrated squared error(AMISE)firstly.Then,particle-driven bandwidth selection is invoked in the KDE.To get a more effective allocation of the particles,the KDE with adap-tive bandwidth in the BAKPF is used to approximate the posterior probability density function(PDF)by moving particles toward the posterior.A closed-form expression of the true distribution is given.The simulation results show that the proposed BAKPF performs better than the standard particle filter(PF),unscented particle filter(UPF)and the kernel particle filter(KPF)both in efficiency and estimation precision.展开更多
In order to improve the performance of the probability hypothesis density(PHD) algorithm based particle filter(PF) in terms of number estimation and states extraction of multiple targets, a new probability hypothesis ...In order to improve the performance of the probability hypothesis density(PHD) algorithm based particle filter(PF) in terms of number estimation and states extraction of multiple targets, a new probability hypothesis density filter algorithm based on marginalized particle and kernel density estimation is proposed, which utilizes the idea of marginalized particle filter to enhance the estimating performance of the PHD. The state variables are decomposed into linear and non-linear parts. The particle filter is adopted to predict and estimate the nonlinear states of multi-target after dimensionality reduction, while the Kalman filter is applied to estimate the linear parts under linear Gaussian condition. Embedding the information of the linear states into the estimated nonlinear states helps to reduce the estimating variance and improve the accuracy of target number estimation. The meanshift kernel density estimation, being of the inherent nature of searching peak value via an adaptive gradient ascent iteration, is introduced to cluster particles and extract target states, which is independent of the target number and can converge to the local peak position of the PHD distribution while avoiding the errors due to the inaccuracy in modeling and parameters estimation. Experiments show that the proposed algorithm can obtain higher tracking accuracy when using fewer sampling particles and is of lower computational complexity compared with the PF-PHD.展开更多
基金Sponsored by Natural Science Foundation of Heilongjiang Province of China(Grant No.QC2001C060)the Science and Technology Research Projectsin Office of Education of Heilongjiang province(Grant No.11531307)
文摘In light of degradation of particle filtering and robust weakness in the utilization of single feature tracking,this paper presents a kernel particle filtering tracking method based on multi-feature integration.In this paper,a new weight upgrading method is given out during kernel particle filtering at first,and then robust tracking is realized by integrating color and texture features under the framework of kernel particle filtering.Space histogram and integral histogram is adopted to calculate color and texture features respectively.These two calculation methods effectively overcome their own defectiveness,and meanwhile,improve the real timing for particle filtering.This algorithm has also improved sampling effectiveness,resolved redundant calculation for particle filtering and degradation of particles.Finally,the experiment for target tracking is realized by using the method under complicated background and shelter.Experiment results show that the method can reliably and accurately track target and deal with target sheltering situation properly.
基金supported by the National Natural Science Foundation of China(60736043,60805012)the Fundamental Research Funds for the Central Universities(K50510020032)
文摘A novel particle filter bandwidth adaption for kernel particle filter(BAKPF)is proposed.Selection of the kernel bandwidth is a critical issue in kernel density estimation(KDE).The plug-in method is adopted to get the global fixed bandwidth by optimizing the asymptotic mean integrated squared error(AMISE)firstly.Then,particle-driven bandwidth selection is invoked in the KDE.To get a more effective allocation of the particles,the KDE with adap-tive bandwidth in the BAKPF is used to approximate the posterior probability density function(PDF)by moving particles toward the posterior.A closed-form expression of the true distribution is given.The simulation results show that the proposed BAKPF performs better than the standard particle filter(PF),unscented particle filter(UPF)and the kernel particle filter(KPF)both in efficiency and estimation precision.
基金Project(61101185) supported by the National Natural Science Foundation of ChinaProject(2011AA1221) supported by the National High Technology Research and Development Program of China
文摘In order to improve the performance of the probability hypothesis density(PHD) algorithm based particle filter(PF) in terms of number estimation and states extraction of multiple targets, a new probability hypothesis density filter algorithm based on marginalized particle and kernel density estimation is proposed, which utilizes the idea of marginalized particle filter to enhance the estimating performance of the PHD. The state variables are decomposed into linear and non-linear parts. The particle filter is adopted to predict and estimate the nonlinear states of multi-target after dimensionality reduction, while the Kalman filter is applied to estimate the linear parts under linear Gaussian condition. Embedding the information of the linear states into the estimated nonlinear states helps to reduce the estimating variance and improve the accuracy of target number estimation. The meanshift kernel density estimation, being of the inherent nature of searching peak value via an adaptive gradient ascent iteration, is introduced to cluster particles and extract target states, which is independent of the target number and can converge to the local peak position of the PHD distribution while avoiding the errors due to the inaccuracy in modeling and parameters estimation. Experiments show that the proposed algorithm can obtain higher tracking accuracy when using fewer sampling particles and is of lower computational complexity compared with the PF-PHD.
