Hello, I'm noticing that theM_RRdot_GPXY matrix created in compute_gp_xy_parameters is transposed with respect to the documentation in SICD 1.5 Vol. 3 pg 66 section (4), where the zero is in the upper right hand side.
Here where m_rrdot_gpxy is defined by a set of nested np.stack() calls:
|
m_rrdot_gpxy = np.negative( |
This chunk of np.stack calls seems to put the zero on the bottom left. Here is the output on a bistatic_small.sicd test file:
866 gp_xy_params = compute_gp_xy_parameters(
867 g[above_threshold, :],
868 ugpn[above_threshold, :],
869 pt_r_rdot_params.bP_PT,
870 pt_r_rdot_params.bPDot_PT,
871 )
(Pdb) pp gp_xy_params.M_RRdot_GPXY
array([[[-8.39392111e-01, -9.28325248e-05],
[-0.00000000e+00, 2.66577045e-02]]])In this case, r_rdot_to_ground_plane_bi() will still converge because the other off-diagonal element is very small, but in other cases it takes many more iterations to converge.
Anyways, this could be intentional depending on how you've organized the system of equations downstream but I figured I would flag it. Thanks!
Hello, I'm noticing that the
M_RRdot_GPXYmatrix created incompute_gp_xy_parametersis transposed with respect to the documentation inSICD 1.5 Vol. 3 pg 66 section (4), where the zero is in the upper right hand side.Here where m_rrdot_gpxy is defined by a set of nested np.stack() calls:
sarkit/sarkit/sicd/projection/_calc.py
Line 1011 in 35aae82
This chunk of np.stack calls seems to put the zero on the bottom left. Here is the output on a bistatic_small.sicd test file:
In this case,
r_rdot_to_ground_plane_bi()will still converge because the other off-diagonal element is very small, but in other cases it takes many more iterations to converge.Anyways, this could be intentional depending on how you've organized the system of equations downstream but I figured I would flag it. Thanks!