kalman_filter

Kalman filter for tracking bounding boxes in image space.

KalmanFilter

 KalmanFilter ()

A Kalman filter class designed for tracking bounding boxes in image space.

Attributes:

ndim (int): The dimension of the state space.
_motion_mat (ndarray): The motion model matrix.
_update_mat (ndarray): The update matrix used for projecting state distribution to measurement space.
_std_weight_position (float): Standard deviation weight for the position.
_std_weight_velocity (float): Standard deviation weight for the velocity.

source

KalmanFilter.init

 KalmanFilter.__init__ ()

Initialize the Kalman filter with default parameters.

source

KalmanFilter._create_std

 KalmanFilter._create_std (mean:numpy.ndarray)

Compute standard deviations based on the mean.

	Type	Details
mean	ndarray	The mean values.
Returns	ndarray	The computed standard deviations.

source

KalmanFilter.initiate

 KalmanFilter.initiate (measurement:numpy.ndarray)

Initialize a new track from an unassociated measurement.

	Type	Details
measurement	ndarray	The initial measurement for the track.
Returns	tuple	The mean and covariance of the initiated track.

source

KalmanFilter.predict

 KalmanFilter.predict (mean:numpy.ndarray, covariance:numpy.ndarray)

Run the Kalman filter prediction step.

	Type	Details
mean	ndarray	The current state mean.
covariance	ndarray	The current state covariance.
Returns	tuple	The predicted state mean and covariance.

source

KalmanFilter.project

 KalmanFilter.project (mean:numpy.ndarray, covariance:numpy.ndarray)

Project the state distribution to the measurement space.

	Type	Details
mean	ndarray	The current state mean.
covariance	ndarray	The current state covariance.
Returns	tuple	The mean and covariance in the measurement space.

source

KalmanFilter.multi_predict

 KalmanFilter.multi_predict (mean:numpy.ndarray, covariance:numpy.ndarray)

Run the Kalman filter prediction step for multiple measurements (Vectorized version).

	Type	Details
mean	ndarray	The current state mean.
covariance	ndarray	The current state covariance.
Returns	tuple	The predicted state mean and covariance for multiple measurements.

source

KalmanFilter.update

 KalmanFilter.update (mean:numpy.ndarray, covariance:numpy.ndarray,
                      measurement:numpy.ndarray)

Run the Kalman filter correction step.

	Type	Details
mean	ndarray	The predicted state mean.
covariance	ndarray	The predicted state covariance.
measurement	ndarray	The new measurement.
Returns	tuple	The updated state mean and covariance after correction.

source

KalmanFilter.gating_distance

 KalmanFilter.gating_distance (mean:numpy.ndarray,
                               covariance:numpy.ndarray,
                               measurements:numpy.ndarray,
                               only_position:bool=False,
                               metric:str='maha')

Compute the gating distance between the state distribution and given measurements.

Raises: ValueError: If an invalid distance metric is provided.

	Type	Default	Details
mean	ndarray		The state mean.
covariance	ndarray		The state covariance.
measurements	ndarray		The given measurements.
only_position	bool	False	If True, consider only position in the gating distance. Defaults to False.
metric	str	maha	The metric to use for distance calculation (‘gaussian’ or ‘maha’). Defaults to ‘maha’.
Returns	ndarray		The gating distances.