states.KalmanBBox

Builds the state-transition matrix ``F``, measurement matrix ``H``, and isotropic noise covariances ``Q``, ``R`` for one of two bbox models: - ``"sort"`` — 7-D state ``[x, y, a, h, vx, vy, va]``, 4-D measurement ``[x, y, a, h]``. Height carries no velocity, matching the original SORT formulation. - ``"deepsort"`` — 8-D state ``[x, y, a, h, vx, vy, va, vh]`` with an added height velocity, matching the canonical Kalman-bbox convention. ``Q`` is parameterised as ``q * I_D`` and treated as per-unit-time process noise: each call accumulates ``q * dt`` along the diagonal of the propagated covariance (see :class:`KalmanLinear` for the exact integration rule). ``R`` is parameterised as ``r * I_4``.