Developing a Real-Time Track Display That Operators Do Not Hate
ABSTRACT We formulate a method of estimating target states that minimizes the mean optimal subpattern assignment (MOSPA) metric, applied suboptimally to a multi-hypothesis tracker (MHT) and optimally to a particle filter. This gives the operator a display of the targets with reduced jitter and track switching.
- SourceAvailable from: nt.ntnu.no[Show abstract] [Hide abstract]
ABSTRACT: For Bayesian filtering of two closely spaced linear Gaussian targets from Gaussian observations, the paper exploits a unique decomposition of the joint conditional density into a mixture of a permutation invariant density and a permutation strictly variant density. This leads to the development of a novel particle filter which performs optimal in the sense of either minimizing track swapping or minimizing track switching, and which includes estimation of the conditional track swap probability. Through Monte Carlo simulations, it is shown that minimizing track switching has a significant advantage over minimizing track swapping, and that the novel particle filter performs remarkably better than a standard particle filter.Decision and Control and European Control Conference (CDC-ECC), 2011 50th IEEE Conference on; 01/2011
- [Show abstract] [Hide abstract]
ABSTRACT: In a preceding paper at Fusion 2009, the existence and characterization of a unique decomposition of the joint conditional density of the states of two targets has been proven. This decomposition consists of a weighted sum of a permutation invariant density and a permutation strictly variant density. In the current paper we exploit this unique decomposition for the development of a novel particle filter for tracking two closely spaced linear Gaussian targets. Thanks to the unique decomposition this novel particle filter is able to provide a conditional estimate of the track swap probability. The remarkable working of this novel particle filter is demonstrated through running Monte Carlo simulations for an example in tracking two closely spaced targets.Information Fusion (FUSION), 2011 Proceedings of the 14th International Conference on; 01/2011
Conference Paper: Finding an OSPA based object detector by aweakly supervised technique[Show abstract] [Hide abstract]
ABSTRACT: The design of multitarget tracking procedures includes, as the most time consuming steps, the definition of the objective class and the formulation of the detection criteria. In this paper we investigate a solution toward an intuitive way for implementing a detector for any ad-hoc application. We capitalize on the OSPA metric to discriminate between the semantic object class of interest and other look-alike classes starting from a short number of unlabeled markers. We propose an illustrative algorithm with a toy example, then we apply it to two real images, the first acquired by SEVIRI, the second by MERIS. In the first case we discriminate between lakes, sea and look-alike clouds, in the other between ground and sea ice. We show how semantic classes with very similar spectral properties can be separated even in the presence of uncertainties or errors in the ground truth.Geoscience and Remote Sensing Symposium (IGARSS), 2012 IEEE International; 01/2012