Page 1
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Bayesian Multiple Target Tracking in Forward Scan Sonar Images Using The PHD Filter
Daniel Clark and Judith Bell
Page 2
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Thesis Topic: “Tomographic Reconstruction of a Sequenceof Forward Scan Sonar Images”
Page 3
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Thesis Topic: “Tomographic Reconstruction of a Sequenceof Forward Scan Sonar Images”
Problems to Address:
Page 4
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Thesis Topic: “Tomographic Reconstruction of a Sequenceof Forward Scan Sonar Images”
Problems to Address:•Segment Sonar into Homogeneous Regions of Same Seabed Type
Page 5
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Thesis Topic: “Tomographic Reconstruction of a Sequenceof Forward Scan Sonar Images”
Problems to Address:•Segment Sonar into Homogeneous Regions of Same Seabed Type•Locate Objects on the Seabed
Page 6
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Thesis Topic: “Tomographic Reconstruction of a Sequenceof Forward Scan Sonar Images”
Problems to Address:•Segment Sonar into Homogeneous Regions of Same Seabed Type•Locate Objects on the Seabed•Align Images onto Global Co-ordinate System
Page 7
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Thesis Topic: “Tomographic Reconstruction of a Sequenceof Forward Scan Sonar Images”
Problems to Address:•Segment Sonar into Homogeneous Regions of Same Seabed Type•Locate Objects on the Seabed•Align Images onto Global Co-ordinate System•Reconstruct Sonar Data into 3D Elevation Map of Seabed
Page 8
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Thesis Topic: “Tomographic Reconstruction of a Sequenceof Forward Scan Sonar Images”
Problems to Address:•Segment Sonar into Homogeneous Regions of Same Seabed Type•Locate Objects on the Seabed•Align Images onto Global Co-ordinate System•Reconstruct Sonar Data into 3D Elevation Map of Seabed
Why?
Page 9
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Thesis Topic: “Tomographic Reconstruction of a Sequenceof Forward Scan Sonar Images”
Problems to Address:•Segment Sonar into Homogeneous Regions of Same Seabed Type•Locate Objects on the Seabed•Align Images onto Global Co-ordinate System•Reconstruct Sonar Data into 3D Elevation Map of Seabed
Why?•To Aid Navigation and Path Planning
Page 10
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Thesis Topic: “Tomographic Reconstruction of a Sequenceof Forward Scan Sonar Images”
Problems to Address:•Segment Sonar into Homogeneous Regions of Same Seabed Type•Locate Objects on the Seabed•Align Images onto Global Co-ordinate System•Reconstruct Sonar Data into 3D Elevation Map of Seabed
Why?•To Aid Navigation and Path Planning•Obstacle Avoidance, mines etc.
Page 11
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Principle of Sonar:Transmission of an acoustic pulse of energy into water andmeasure reflected energy:
Page 12
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Principle of Sonar:Transmission of an acoustic pulse of energy into water andmeasure reflected energy:
The intensity of the energy reflected is measured against time to give information on the surface below:
distance to surface = (speed of sound in water)x(time to return)/2
Page 13
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Forward Scan Sonar:•The acoustic energy from the sonar is transmitted in a radial sector.
Page 14
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Forward Scan Sonar:•The acoustic energy from the sonar is transmitted in a radial sector.•The backscattered energy can be shown as a sonar image:
Page 15
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Environmental ReconstructionMethods for creating elevation maps of the seabed have been implemented using Lambert's Law and knowledge of the Sonar's Position.
Page 16
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Environmental ReconstructionMethods for creating elevation maps of the seabed have been implemented using Lambert's Law and knowledge of the Sonar's Position.
Lambert's Law relates the reflected energy from the surface to the anglebetween the direction of reflection and the surface normal.
I s I I cos2
Page 17
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Local Propagation Technique:Each successive point is determined from the intersection ofthe circle centred at the sonar fish and the surface gradientdetermined from Lambert's Law.
Page 18
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Sonar Image:
Page 19
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Sonar Image:
Reconstructed Image:
Page 20
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Multiple Target Tracking
This work will be used to address the issues of:•Detecting and Locating Objects on the Seabed•Aligning Images onto a Global Co-ordinate System
Page 21
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Multiple Target Tracking
This work will be used to address the issues of:•Detecting and Locating Objects on the Seabed•Aligning Images onto a Global Co-ordinate System
Targets to be Tracked:Mines, metallic objects which have high reflectance property:
Measurements obtained by thresholding.
Page 22
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Recursive Bayesian Estimation
To make inference about a dynamic system, two models are needed:
•Motion Model – describes evolution of state with time ie the motion of underwater vehicle.
Page 23
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Recursive Bayesian Estimation
To make inference about a dynamic system, two models are needed:
•Motion Model – describes evolution of state with time ie the motion of underwater vehicle.
•Measurement Model – relates the measurements to the state ie the objects on the seabed.
Page 24
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Recursive Bayesian Estimation
To make inference about a dynamic system, two models are needed:
•Motion Model – describes evolution of state with time ie the motion of underwater vehicle.
•Measurement Model – relates the measurements to the state ie the objects on the seabed.
These correspond to prediction and update stages when tracking.
Page 25
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Single Target Inference
The tracking problem is governed by two functions:
These relate to the motion and measurement models respectively.
x t F t x t 1 , v t 1
zt H t x t , nt
Page 26
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Single Target Inference
The tracking problem is governed by two functions:
These relate to the motion and measurement models respectively.•The process noise v reflects the unknown target motion •The measurement noise n reflects sensor errors
x t F t x t 1 , v t 1
zt H t x t , nt
Page 27
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Bayesian Recursion
The prior distribution of a target location based on previous observationsis obtained from the motion model and the posterior at time t-1:
f t t 1 x t z1 :t 1 f t t 1 x t x t 1 f t 1 t 1 x t 1 z1: t 1 dx t 1
Page 28
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Bayesian Recursion
The prior distribution of a target location based on previous observationsis obtained from the motion model and the posterior at time t-1:
When a new measurement is obtained, the posterior distribution at time tis obtained by Bayes' Law:
where g is the likelihood of observing z given target state x.
f t t 1 x t z1 :t 1 f t t 1 x t x t 1 f t 1 t 1 x t 1 z1: t 1 dx t 1
f t t x t z1. .t g t zt x t f t t 1 x t z1. .t 1
Page 29
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Multiple Target Inference Model
The single target recursive state estimation can be directly extended to a multiple target model using Finite Set Statistics (FISST).
Page 30
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Multiple Target Inference Model
The single target recursive state estimation can be directly extended to a multiple target model using Finite Set Statistics (FISST).
A Random Finite Set (RFS) is used to represent a multiple-target state.The set of objects tracked at time t is an RFS containing :•Set of objects survived from time t-1.
Page 31
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Multiple Target Inference Model
The single target recursive state estimation can be directly extended to a multiple target model using Finite Set Statistics (FISST).
A Random Finite Set (RFS) is used to represent a multiple-target state.The set of objects tracked at time t is an RFS containing :•Set of objects survived from time t-1.•Set of object appearing at time t.
Page 32
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Multiple Target Inference Model
The single target recursive state estimation can be directly extended to a multiple target model using Finite Set Statistics (FISST).
A Random Finite Set (RFS) is used to represent a multiple-target state.The set of objects tracked at time t is an RFS containing :•Set of objects survived from time t-1.•Set of object appearing at time t.
The measurements at time t are modelled by an RFS containing:•Measurements generated from actual targets.
Page 33
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Multiple Target Inference Model
The single target recursive state estimation can be directly extended to a multiple target model using Finite Set Statistics (FISST).
A Random Finite Set (RFS) is used to represent a multiple-target state.The set of objects tracked at time t is an RFS containing :•Set of objects survived from time t-1.•Set of object appearing at time t.
The measurements at time t are modelled by an RFS containing:•Measurements generated from actual targets.•Spurious measurements from clutter.
Page 34
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Multiple Target Bayesian Recursion
The single target Bayesian recursion can be extended to the multipletarget scenario using the calculus defined in FISST:
f t t 1 X t Z1 :t 1 f t t 1 X t X t 1 , Z1 :t 1 f t 1 t 1 X t 1 Z1 : t 1 X t 1
f t t X t Z 1: t gt Z t X t f t t 1 X t Z1:t 1
Page 35
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
PHD Filter
The Probability Hypothesis Density or PHD is defined as the first orderstatistical moment of the multiple target posterior distribution.The integral of the PHD in any region represents the expected numberof objects in that region.
Page 36
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
PHD Filter
The Probability Hypothesis Density or PHD is defined as the first orderstatistical moment of the multiple target posterior distribution.The integral of the PHD in any region represents the expected numberof objects in that region.
Why bother?Computationally cheaper to calculate than full posterior.
Page 37
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
PHD Filter
The Probability Hypothesis Density or PHD is defined as the first orderstatistical moment of the multiple target posterior distribution.The integral of the PHD in any region represents the expected numberof objects in that region.
Why bother?Computationally cheaper to calculate than full posterior.
Where are the targets?The locations of the targets can be found as peaks of this distribution.
Page 38
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
PHD Filter
Prediction Equation:
Dt t 1 x t Z1:t 1 b t x t PS x t 1 f t t 1 x t x t 1 Dt 1 t 1 x t 1 Z1 :t 1 x t 1
Page 39
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
PHD Filter
Prediction Equation:
Data Update Equation:
Dt t 1 x t Z1:t 1 b t x t PS x t 1 f t t 1 x t x t 1 Dt 1 t 1 x t 1 Z1 :t 1 x t 1
Dt t x t Z1:t F t Z t x t Dt t 1 x t Z1:t 1
F t Z t x t 1 PD zti Z t
PD gt zti x t
t c t zt PD Dt t , g t
Page 40
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Particle Filter Algorithm
Particle filters were designed for implementing Bayesian recursionby representing probability distributions by random samples or particles rather than in their functional forms.
Page 41
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Particle Filter Algorithm
Particle filters were designed for implementing Bayesian recursionby representing probability distributions by random samples or particles rather than in their functional forms.
The areas with higher probability will have a larger number of particles.
Page 42
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Particle Filter Algorithm (Single Target)
•Step 1: Initialisation.N particles are uniformly distributed across the Field of View (FoV).
Page 43
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Particle Filter Algorithm (Single Target)
•Step 1: Initialisation.N particles are uniformly distributed across the Field of View (FoV).
•Step 2: Data Update.After a new measurement, weights are assigned to the particlesaccording to their likelihoods.
Page 44
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Particle Filter Algorithm (Single Target)
•Step 1: Initialisation.N particles are uniformly distributed across the Field of View (FoV).
•Step 2: Data Update.After a new measurement, weights are assigned to the particlesaccording to their likelihoods.
•Step 3: ResamplingAn unweighted particle set is obtained by resampling from the weighted set.
Page 45
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Particle Filter Algorithm (Single Target)
•Step 1: Initialisation.N particles are uniformly distributed across the Field of View (FoV).
•Step 2: Data Update.After a new measurement, weights are assigned to the particlesaccording to their likelihoods.
•Step 3: ResamplingAn unweighted particle set is obtained by resampling from the weighted set.
•Step 4: Estimation of Target LocationThe location of the target is estimated by calculating the mean positionof the particles.
Page 46
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Particle Filter Algorithm (Single Target)
•Step 1: Initialisation.N particles are uniformly distributed across the Field of View (FoV).
•Step 2: Data Update.After a new measurement, weights are assigned to the particlesaccording to their likelihoods.
•Step 3: ResamplingAn unweighted particle set is obtained by resampling from the weighted set.
•Step 4: Estimation of Target LocationThe location of the target is estimated by calculating the mean positionof the particles.
•Step 5: PredictionThe location of the next target location is estimated using the motion model.
Page 47
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Particle Filter Algorithm (Multiple Target)
•Step 1: Initialisation.N particles are uniformly distributed across the Field of View (FoV).
Page 48
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Particle Filter Algorithm (Multiple Target)
•Step 1: Initialisation.N particles are uniformly distributed across the Field of View (FoV).
•Step 2: Data Update.After a new measurement, weights are assigned to the particlesaccording to the data update equation for the PHD.
Page 49
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Particle Filter Algorithm (Multiple Target)
•Step 1: Initialisation.N particles are uniformly distributed across the Field of View (FoV).
•Step 2: Data Update.After a new measurement, weights are assigned to the particlesaccording to the data update equation for the PHD.
•Step 3: Estimation of Target LocationThe locations of the targets are estimated by fitting a Gaussianmixture model to the particles where the number of targets is the sum of the weights.
Page 50
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Particle Filter Algorithm (Multiple Target)
•Step 1: Initialisation.N particles are uniformly distributed across the Field of View (FoV).
•Step 2: Data Update.After a new measurement, weights are assigned to the particlesaccording to the data update equation for the PHD.
•Step 3: Estimation of Target LocationThe locations of the targets are estimated by fitting a Gaussianmixture model to the particles where the number of targets is the sum of the weights.
•Step 4: ResamplingAn unweighted particle set is obtained by resampling from the weighted set.
Page 51
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Particle Filter Algorithm (Multiple Target)
•Step 1: Initialisation.N particles are uniformly distributed across the Field of View (FoV).
•Step 2: Data Update.After a new measurement, weights are assigned to the particlesaccording to the data update equation for the PHD.
•Step 3: Estimation of Target LocationThe locations of the targets are estimated by fitting a Gaussianmixture model to the particles where the number of targets is the sum of the weights.
•Step 4: ResamplingAn unweighted particle set is obtained by resampling from the weighted set.
•Step 5: PredictionThe locations of the next target locations are estimated using the PHDprediction equation.
Page 52
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Implementation on Forward Scan Sonar Data
•n beams separated by k degrees
Page 53
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Implementation on Forward Scan Sonar Data
•n beams separated by k degrees•Intensity vs time data
Page 54
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Implementation on Forward Scan Sonar Data
•n beams separated by k degrees•Intensity vs time data•Tracks range and bearing of targets
Page 55
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Implementation on Forward Scan Sonar Data
•n beams separated by k degrees•Intensity vs time data•Tracks range and bearing of targets
Measurements are obtained by thresholding the data:
Page 56
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Implementation on Forward Scan Sonar Data
•n beams separated by k degrees•Intensity vs time data•Tracks range and bearing of targets
Measurements are obtained by thresholding the data:
The motion of the sonar is assumed to be linear Field of View in the range of 20m to 60m
Page 57
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Implementation on Forward Scan Sonar Data
Sequence of Simulated Forward Scan Sonar Images with Objects
Page 58
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Results on Simulated Data
Linear Tracking in Sonar Image Reference Frame
Page 59
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Results on Simulated Data
Linear Tracking in Global Reference Frame
Page 60
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Results on Simulated Data
Sinusoidal Tracking in Global Reference Frame
Page 61
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Implementation on Forward Scan Sonar Data
Sequence of Real Forward-Scan Images
Page 62
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Results on Real Data
Tracked Cylinder in Forward Direction
Page 63
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Results on Real Data
Tracked Cylinder in Backward Direction
Page 64
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Extensions To PHD Filter:
The multi-target state set does not give individual target identities.
How do we associate measurements between frames?
Two possible approaches:
Page 65
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Extensions To PHD Filter:
The multi-target state set does not give individual target identities.
How do we associate measurements between frames?
Two possible approaches:
•Increase state vector with invariant attribute
Page 66
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Extensions To PHD Filter:
The multi-target state set does not give individual target identities.
How do we associate measurements between frames?
Two possible approaches:
•Increase state vector with invariant attribute
•Data Association
Page 67
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
What do I do with it now?
The target tracks will be used for alignment by computinga geometric transform between the frames eg affine/similarity.
Page 68
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
What do I do with it now?
The target tracks will be used for alignment by computinga geometric transform between the frames eg affine/similarity.
This will be used as a preliminary step for 3D reconstruction of asequence of images.
Page 69
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
Future Work on Reconstruction
Segmentation into different seabed types.
Reconstructed 3D elevation map of sequence of images.
Page 70
Forward-Scan Sonar Tomographic Forward-Scan Sonar Tomographic Reconstruction Reconstruction
PHD Filter Multiple Target TrackingPHD Filter Multiple Target Tracking
[email protected]
References
R. Mahler:"Multitarget Bayes Filtering via First-Order Multitarget Moments",IEEE Transactions on Aerospace and Electronic Systems, 2003.
Vo, Singh and Doucet:“Sequential Monte Carlo Implementation of the PHD Filter for Multi-targetTracking”Proc. FUSION 2003
Sidenbladh and Wirkander:“Tracking random sets of vehicles in terrain”IEEE Workshop on Multi-Object Tracking, Madison, WI, USA, 2003