[RSDK-13312] Fixing order of parameters for Brown Conrady K6

[SebastianMunozP]

This PR fixes the order of parameters for Brown Conrady K6 distortion model used in Astra2 for the get_properties call

The counterpart to this fix, to actually support such distortion model is in this PR: viamrobotics/rdk#5702

The order of parameters that the RDK is as follows:

            type BrownConradyK6 struct {
            RadialK1     float64 `json:"rk1"`
            RadialK2     float64 `json:"rk2"`
            RadialK3     float64 `json:"rk3"`
            RadialK4     float64 `json:"rk4"`
            RadialK5     float64 `json:"rk5"`
            RadialK6     float64 `json:"rk6"`
            TangentialP1 float64 `json:"tp1"`
            TangentialP2 float64 `json:"tp2"`
            }

Testing

With this test we confirm that the order of parameters is correct:
Parameters sent by Orbbec (taken from logs below):

0.114636, //K1
-0.328871, //K2
0.256854, //K3
0, //K4
0, //K5
0, //K6
-0.000840949, //P1
9.80989e-05//P2

Parameters received by Python client (taken from logs below)

    "distortion": {
      "model": "BROWN_CONRADY_K6",
      "k1": 0.11463635414838791,
      "k2": -0.3288709819316864
      "k3": 0.2568541169166565,,
      "k4": 0,
      "k5": 0
      "k6": 0,
      "p1": -0.0008409488946199417,
      "p2": 0.00009809886250877753,
    },

Logs

Orbbec Get Properties Log:

2/2/2026, 2:38:47 PM debug orbbec-1     [module/orbbec.cpp:1273] [get_properties] end   log_ts 2026-02-02T14:38:47.570Z

2/2/2026, 2:38:47 PM debug orbbec-1     [module/orbbec.cpp:1271] [get_properties] Properties: distortion parameters: [model: BROWN_CONRADY_K6, parameters: [0.114636, -0.328871, 0.256854, 0, 0, 0, -0.000840949, 9.80989e-05, ]], intrinsic parameters: [1280, 720, 836.655, 837.041, 640.232, 351.624]   log_ts 2026-02-02T14:38:47.570Z

2/2/2026, 2:38:47 PM debug orbbec-1     [module/orbbec.cpp:1195] [get_properties] start   log_ts 2026-02-02T14:38:47.570Z

Getting distortion parameters through DoCommand confirms that the order of parameters is correct

{
  "Color": {
    "intrinsics": {
      "fx": 836.6552124023438,
      "cx": 640.2318725585938,
      "cy": 351.623779296875,
      "width": 1280,
      "fy": 837.0411987304688,
      "height": 720
    },
    "height": 720,
    "width": 1280,
    "distortion": {
      "p1": -0.0008409488946199417,
      "k6": 0,
      "k4": 0,
      "k3": 0.2568541169166565,
      "k1": 0.11463635414838791,
      "k2": -0.3288709819316864,
      "model": "BROWN_CONRADY_K6",
      "p2": 0.00009809886250877753,
      "k5": 0
    },
    "format": "MJPG",
    "fps": 30
  },
  "Depth": {
    "width": 1600,
    "distortion": {
      "k3": 0,
      "k6": 0,
      "k2": 0,
      "model": "BROWN_CONRADY_K6",
      "k1": 0,
      "k5": 0,
      "p1": 0,
      "k4": 0,
      "p2": 0
    },
    "format": "Y16",
    "intrinsics": {
      "cx": 800.1470336914062,
      "fy": 1424.7021484375,
      "height": 1200,
      "fx": 1424.67822265625,
      "width": 1600,
      "cy": 596.7117919921875
    },
    "fps": 30,
    "height": 1200
  }
}

[seanavery]

LGTM. Are we planning to use Brown Conrady distortion model for camera calibration in a surface finishing or other use case?

[SebastianMunozP]

LGTM. Are we planning to use Brown Conrady distortion model for camera calibration in a surface finishing or other use case?

Not really, this is mainly to fix the get_properties call that was broken in Orbbec, to circumvent this we were using a do_command, so this is just tech debt.