getStereoPairs() returns empty on OAK-D (BW1098OBC, rev R0M0E0) with depthai 2.32.0 — "Didn't find any boards matching the eeprom data" #1127

Hello Luxonis support team,

I am using an OAK-D camera with the depthai 2.32.0.0 Python package on Linux (Python 3.14) and I cannot get any stereo pairs from the device, although the camera and its calibration are working correctly.

Device information

• Product name: OAK-D
• MX ID: 14442C10A148C1D200
• Board name (from EEPROM): BW1098OBC
• Board revision (from EEPROM): R0M0E0
• Platform: X_LINK_MYRIAD_X, USB VSC, USB 3.0 (SUPER)
• Connected cameras:
  • CAM_A: IMX378 (color, "color")
  • CAM_B: OV9282 (mono, "left")
  • CAM_C: OV9282 (mono, "right")

Problem

Both Device.getStereoPairs() and Device.getAvailableStereoPairs() return empty lists, even though three cameras are connected and the EEPROM is readable.

Evidence (from DEPTHAI_DEBUG=1 DEPTHAI_LEVEL=debug)

1. The host-side board config is loaded with an empty camera array:

   [host] [debug] Device - BoardConfig: {"camera":[],"emmc":null,"gpio":[],"imu":null,...}
   

2. The device firmware reports that no board matches the EEPROM data (printed twice during boot):

   [system] [warning] PRINT:LeonCss: Didn't find any boards matching the eeprom data
   

3. The EEPROM calibration itself is intact and contains a complete stereo setup:

   stereoRectificationData.leftCameraSocket  = 1  (CAM_B)
   stereoRectificationData.rightCameraSocket = 2  (CAM_C)
   baseline                                  = 7.5 mm
   stereoUseSpecTranslation                  = true
   

   CalibrationHandler.getStereoLeftCameraId() → CAM_B
CalibrationHandler.getStereoRightCameraId() → CAM_C
CalibrationHandler.getBaselineDistance() → 7.5 mm
   Rectification rotation matrices for both cameras are present and valid.

Questions

1. Is the empty result of getStereoPairs() a known issue for board BW1098OBC revision R0M0E0 in depthai 2.32.0 (i.e. the board is missing from or mismatched in the board configuration database)?
2. Is there a fixed or newer depthai release that resolves this?
3. If the board revision is not in the database, is there a way to provide a custom board configuration (e.g. via CalibrationHandler(calibrationDataPath, boardConfigPath) or a board config JSON) so that getStereoPairs() works?

Workaround currently in use

Since the EEPROM calibration is complete, I derive the stereo setup from CalibrationHandler (left/right camera IDs, baseline, intrinsics, rectification rotations) instead of relying on getStereoPairs(). This works, but I would like to confirm whether the empty getStereoPairs() result is expected behavior or a bug.

Thank you for your help.

Best regards

debug_log.txt
eeprom_data.json