Poor Depth Quality on OAK FFC 4P with OV9782 Stereo Pair

ShreyasK

I am using a Luxonis OAK-FFC-4P module with OV9782 W sensors connected at B and C ports. My goal is to generate a depth map from this stereo pair. I will be using the default un-distortion method through enableUndistortion to reduce the FOV and get undistorted image feed. Hoping this extra layer of processing will not bother a much for the depth calculations.

I followed the official documentation on Configuring Stereo Depth and came up with the code attached.

Additional Information:

Stereo baseline: 5 cm (same as Intel RealSense baseline)
Distance from object: ~45 cm
Use cases: VIO, obstacle avoidance, and other drone applications
Observed issue: Disparity map from OAK is noisy and inaccurate, unlike the D435i output.

However, the depth output quality is far from expected. To validate this, I ran a comparison against an Intel RealSense D435i connected to the same PC, with both devices placed at the same distance (approx. 45 cm from a water bottle used as the target object). For a fair comparison, I disabled the infrared emitter on the D435i.

The depth results from the RealSense are significantly more reliable, at least visually and stable, than those from the OAK module. I’ve attached images of the depth maps from both devices for reference.

Depth map from OAK Module:

Infra Feed from OAK Module:

Realsense Depth Map:

Code:

#!/usr/bin/env python3

import cv2

import depthai as dai

import numpy as np

pipeline = dai.Pipeline()

monoLeft = pipeline.create(dai.node.Camera).build(dai.CameraBoardSocket.CAM_B)

monoRight = pipeline.create(dai.node.Camera).build(dai.CameraBoardSocket.CAM_C)

stereo = pipeline.create(dai.node.StereoDepth)

# Linking

monoLeft.initialControl.setSharpness(2)

monoLeft.initialControl.setLumaDenoise(1)

monoLeft.initialControl.setChromaDenoise(4)

monoLeft.initialControl.setAntiBandingMode(dai.CameraControl.AntiBandingMode.MAINS_50_HZ)

monoRight.initialControl.setSharpness(2)

monoRight.initialControl.setLumaDenoise(1)

monoRight.initialControl.setChromaDenoise(4)

monoRight.initialControl.setAntiBandingMode(dai.CameraControl.AntiBandingMode.MAINS_50_HZ)

monoLeftOut = monoLeft.requestOutput(

size=(640,400),

type=dai.ImgFrame.Type.NV12,

resizeMode=dai.ImgResizeMode.CROP,

enableUndistortion=True,

fps=30

)

monoRightOut = monoRight.requestOutput(

size=(640,400),

type=dai.ImgFrame.Type.NV12,

resizeMode=dai.ImgResizeMode.CROP,

enableUndistortion=True,

fps=30

)

monoLeftOut.link(stereo.left)

monoRightOut.link(stereo.right)

stereo.setRectification(True)

stereo.setExtendedDisparity(True)

stereo.setLeftRightCheck(True)

stereo.setSubpixel(False)

stereo.setSubpixelFractionalBits(4)

# stereo.initialConfig.setConfidenceThreshold(0)

stereo.PresetMode(5)

stereo.setBaseline(5)

stereo.setPostProcessingHardwareResources(3, 3)

stereo.initialConfig.setMedianFilter(dai.StereoDepthConfig.MedianFilter.KERNEL_3x3)

stereo.initialConfig.postProcessing.temporalFilter.enable = True

stereo.initialConfig.postProcessing.temporalFilter.alpha = 0.5

stereo.initialConfig.postProcessing.temporalFilter.delta = 3

stereo.initialConfig.postProcessing.speckleFilter.enable = True

stereo.initialConfig.postProcessing.speckleFilter.speckleRange = 12

stereo.initialConfig.postProcessing.spatialFilter.enable = True

stereo.initialConfig.postProcessing.spatialFilter.alpha = 0.5

stereo.initialConfig.postProcessing.spatialFilter.delta = 3

stereo.initialConfig.postProcessing.spatialFilter.holeFillingRadius = 2

stereo.initialConfig.postProcessing.spatialFilter.numIterations = 1

stereo.initialConfig.postProcessing.decimationFilter.decimationFactor = 2

# stereo.initialConfig.postProcessing.thresholdFilter.minRange = 400

# stereo.initialConfig.postProcessing.thresholdFilter.maxRange = 15000

syncedLeftQueue = stereo.syncedLeft.createOutputQueue()

syncedRightQueue = stereo.syncedRight.createOutputQueue()

disparityQueue = stereo.disparity.createOutputQueue()

colorMap = cv2.applyColorMap(np.arange(256, dtype=np.uint8), cv2.COLORMAP_JET)

colorMap[0] = [0, 0, 0] # to make zero-disparity pixels black

with pipeline:

pipeline.start()

maxDisparity = 1

while pipeline.isRunning():

leftSynced = syncedLeftQueue.get()

rightSynced = syncedRightQueue.get()

disparity = disparityQueue.get()

assert isinstance(leftSynced, dai.ImgFrame)

assert isinstance(rightSynced, dai.ImgFrame)

assert isinstance(disparity, dai.ImgFrame)

cv2.imshow("left", leftSynced.getCvFrame())

cv2.imshow("right", rightSynced.getCvFrame())

npDisparity = disparity.getFrame()

maxDisparity = max(maxDisparity, np.max(npDisparity))

# Display disparity as greyscale instead of color

greyscaleDisparity = ((npDisparity / maxDisparity) * 255).astype(np.uint8)

cv2.imshow("disparity", greyscaleDisparity)

key = cv2.waitKey(1)

if key == ord('q'):

pipeline.stop()

break

Questions for the Luxonis Team

Is there any sensor-specific tuning required for OV9782 W modules to improve stereo depth quality?
Is the 5 cm baseline setup fully supported for stereo depth on the OAK FFC 4P?
Are there additional calibration or parameter tweaks (e.g., Subpixel mode, confidence threshold, median filter, or other advanced configs) that I should try?
Is there a recommended pipeline template for getting the best stereo performance from OV9782 sensors on OAK FFC 4P?
Is the current performance limitation hardware-related (sensor choice, baseline, etc.) or software/tuning-related?
When subpixel is turned with median filter, it ends up with below errors:

[18443010313351F500] [3.4] [4.993] [StereoDepth(2)] [error] Maximum disparity value '3040' exceeds the maximum supported '1024' by median filter. Disabling median filter!

[18443010313351F500] [3.4] [5.046] [StereoDepth(2)] [error] Maximum disparity value '3040' exceeds the maximum supported '1024' by median filter. Disabling median filter!

This issue is opened to understand if I’ve missed any configuration steps or if there are known limitations with OV9782 stereo modules compared to Realsense.

Looking forward to any guidance, best practices, or sample configurations from the Luxonis team.

Oskar-Luxonis

Hi @ShreyasK thanks for the detailed report!

Main issue I see is that you’re undistorting in the Camera node and letting StereoDepth rectify, which double-warps the images and degrades matching. I’d suggest starting from our v3 StereoDepth example and adjusting from there:

https://docs.luxonis.com/software-v3/depthai/examples/stereo_depth/stereo_depth/

Also, did you calibrate the cameras? Guide here:
https://docs.luxonis.com/hardware/platform/depth/calibration/

Hope it helps,
Oskar

ShreyasK

Thank you for reply @Oskar-Luxonis

I will attach few images for your kind reference:

"Bottle" is placed at a distance of 44cm approx.

This the default stereo.py
This default stereo.py with resolution of 640,400.
This actual camera feed:

Calibration is done this is the output:

After Calibration:

640,400 with default code:
default code(FUllresolution output)
Camera Feed(640,400):

Based on these, I felt calibration improved the stuff majorly. And undistorted feed cannot be used for Stereo.

So few question:

do u have any suggestions on setting etc (I have read configuring stereo depth page but need little info from personal experience with Wide-angle lens and oak FFC module).
So as I am working on a project where it needs undistorted feed for VIO and stereo depth for obstacle avoidance, how can I get two feed at a time. Because undistortion setting can be turned on only in requestOutput?
Or any trick to use undistorted feed with stereo depth?
Left camera feels more blurry at corners (I do not mean extreme corner). Tried cleaning etc but didn't able to fix the same.

jakaskerl

ShreyasK
Looks like like a bad calibration.. Can you send the configuration.json you used during calibration and the command you used to run it?

Thanks
Jaka

ShreyasK

Sure @jakaskerl.

configuration.json file:

I used 32 inch charuco pdf on my 27 inch.

Distance between cameras is 5cm.

And this is the command I used:
python3 calibrate.py -s 3.284 -nx 17 -ny 9 -brd OAK-FFC-4P.json

jakaskerl

ShreyasK
Try specifying this for both stereo cameras:

monoRight = pipeline.create(dai.node.Camera).build(dai.CameraBoardSocket.CAM_C)
monoRight.setSensorType(dai.CameraSensorType.COLOR)

This should force the board to treat OV9782 as a COLOR sensor so it doesn't 2x2 bin it to 400p and break the image.

Thanks,
Jaka

ShreyasK

@jakaskerl Will try the same and check performance.

Is the calibration and stuff correct?

jakaskerl

ShreyasK
Yeah looks fine. Only issue I can see would be the wrong placement of sensors (mix-up of L and R).

Thanks,
Jaka

ShreyasK

Do u mean that right should be left and left should be right?
The left and right configuration is correct, I have rechecked it.
This is the output of using this: monoRight.setSensorType(dai.CameraSensorType.COLOR)

Compared to default mono:
With few filters like and also updated to newer stable release 3.0.0:
stereo.setSubpixel(True)

stereo.setSubpixelFractionalBits(4)

stereo.PresetMode(dai.node.StereoDepth.PresetMode.ROBOTICS)

stereo.setPostProcessingHardwareResources(3, 3)

stereo.initialConfig.postProcessing.temporalFilter.enable = True

stereo.initialConfig.postProcessing.temporalFilter.alpha = 0.5

stereo.initialConfig.postProcessing.temporalFilter.delta = 3

stereo.initialConfig.postProcessing.speckleFilter.enable = True

stereo.initialConfig.postProcessing.speckleFilter.speckleRange = 5

stereo.initialConfig.postProcessing.spatialFilter.enable = True

stereo.initialConfig.postProcessing.spatialFilter.alpha = 0.5

stereo.initialConfig.postProcessing.spatialFilter.delta = 3

stereo.initialConfig.postProcessing.spatialFilter.holeFillingRadius = 2

stereo.initialConfig.postProcessing.spatialFilter.numIterations = 1

stereo.initialConfig.postProcessing.decimationFilter.decimationFactor = 2

Output
This is the output of real sense from same spot the oak camera was kept:
Is there any way we can get output from decimation filter in desired resolution than that short output?

ShreyasK

Hi @jakaskerl
Any inputs from your end?

jakaskerl

ShreyasK
Was on vacation, apologies for late reply. The image looks good, I was mainly concerned with the 400p image since binning is done when downsizing from 800p.

If you want better depth, you need to improve the texture of the surfaces you are looking at. The bottle looks completely black so the algorithm can not perform depthai.

Thanks,
Jaka

ShreyasK

jakaskerl Thank you for your reply.

With regard to binning, as I need grayscale than color, will there be any way on this regard?
Because if I set sensorType to color it give color output which will affect the other systems which need the greyscale image feed.

ShreyasK

Oskar-Luxonis Thank you so much for your quick reply.
So now in my realworld usecase for VIO and obstacle avoidance i need 640*400 image output from stereo pair in mono/grayscale.
when use 1280x800 and color output but process it with imagemanip node and convert it to (640x400, mono) for my usecase. doesnt that add extra load on oak ffc4p module leading to issues in latency or delay as it already grabs and runs 4 cam and also stereo and aligndepth(External power supply other than usb is given). as realtimeness too is one of the factor i need to be concerned for VIO.
With regard to stereo depth, I can go ahead with decimationfilter=2 to get 640x400?
Using direct 640x400 is really a concern? like does it complicate things?
Previous conversation @jakaskerl informed about mono and resolution realted stuff.

jakaskerl I was mainly concerned with the 400p image since binning is done when downsizing from 800p.

jakaskerl

Hi ShreyasK
The OV9728 sensor is a color sensor. If it's shows a mono stream then it's already been converted. You can manually convert it to mono using imageManip if you really need to - otherwise, there shouldn't be any issues just using the color stream.

Thanks,
Jaka

ShreyasK

Hi @jakaskerl & @Oskar-Luxonis
I want to have review on newer calibration values with clarity on whats the right way to apply rotation on the correct camera.
I have 4 camera connected to oak ffc4p. 4 cameras are OAK FFC OV9782 FF. 3 are facing front and 1 camera is mounted down facing. I need majorly calibration for front 3 camera, and 4th camera facing down data is just added so there exists some calibration data.
Now the major issue here is, stereo pair cam_b and cam_c are normally placed with baseline of 6cm. but cam_a is placed 180 degrees rotated in the middle; 3cm far from cam_b and cam_c. As all three cams need to be calibrated for align depth i went ahead with below json. can u please help me correct this if any mistakes are there in the json.
I used dsb param for disabling 4th camera when 3 cameras are calibrated, then i did calibration of just 4th camera disabling other 3.

{
    "board_config": {
        "name": "OAK-FFC-4P",
        "cameras": {
            "CAM_B": {
                "name": "left",
                "hfov": 80,
                "type": "mono"
            },
            "CAM_A": {
                "name": "middle",
                "hfov": 80,
                "type": "color",
                "extrinsics": {
                    "to_cam": "CAM_B",
                    "specTranslation": {
                        "x": 3,
                        "y": 0,
                        "z": 0
                    },
                    "rotation": {
                        "r": 0,
                        "p": 0,
                        "y": 180
                    }
                }
            },
            "CAM_C": {
                "name": "right",
                "hfov": 80,
                "type": "mono",
                "extrinsics": {
                    "to_cam": "CAM_B",
                    "specTranslation": {
                        "x": 6,
                        "y": 0,
                        "z": 0
                    },
                    "rotation": {
                        "r": 0,
                        "p": 0,
                        "y": 0
                    }
                }
            },
            "CAM_D": {
                "name": "down",
                "hfov": 80,
                "type": "mono",
                "extrinsics": {
                    "to_cam": "CAM_B",
                    "specTranslation": {
                        "x": 10,
                        "y": 0,
                        "z": 0
                    },
                    "rotation": {
                        "r": 0,
                        "p": 0,
                        "y": 0
                    }
                }
            }
        },
        "stereo_config": {
            "left_cam": "CAM_B",
            "right_cam": "CAM_C"
        }
    }
}

I have refered to preset board config files but i wasn't able to find a single file applying rotation.
This below cli output is rerun of calibration for just stereo pair:

`<-------------Extrinsics calibration of right and left ------------> Use DEFAULT model`

`Epipolar error is 0.23235119675837798`

`Printing Extrinsics res... [[ 0.99997247 -0.00708394  0.00220968] [ 0.00710053  0.99994594 -0.00759604] [-0.00215575  0.00761152  0.99996871]] [[ 5.81244231e+00] [-2.03211534e-03] [ 6.24935640e-03]] Euler angles in XYZ [0.43611321 0.12351542 0.40683502] degs <-------------Epipolar error of right and left ------------> Epipolar error 0.23235119675837798`

`Reprojection error threshold -> 1.1111111111111112`

`left Reprojection Error: 0.227620`

`Reprojection error threshold -> 1.1111111111111112`

`right Reprojection Error: 0.214708`

`Flashing Calibration data into EEPROM VERSION being flashed is  -> 7`

`EEPROM VERSION being flashed is  -> 7`

`py: DONE.`

And this is output of calibration_dump:

Is EEPROM available: True
User calibration: {
  "batchName": "",
  "batchTime": 1717298050,
  "boardConf": "IR-C00M00-00",
  "boardCustom": "",
  "boardName": "DD2090",
  "boardOptions": 0,
  "boardRev": "R7M1E7",
  "cameraData": [
    [
      0,
      {
        "cameraType": 0,
        "distortionCoeff": [
          4.428467273712158,
          12.631766319274902,
          -0.0008713753195479512,
          -0.0005999229615554214,
          -24.348819732666016,
          4.171751022338867,
          13.42512321472168,
          -24.945770263671875,
          0.0,
          0.0,
          0.0,
          0.0,
          0.0,
          0.0
        ],
        "extrinsics": {
          "rotationMatrix": [
            [
              -0.999979555606842,
              0.0026230337098240852,
              0.005833887029439211
            ],
            [
              -0.0026709451340138912,
              -0.9999626278877258,
              -0.008220034651458263
            ],
            [
              0.005812107585370541,
              -0.008235448971390724,
              0.9999492168426514
            ]
          ],
          "specTranslation": {
            "x": 3.0,
            "y": 0.0,
            "z": 0.0
          },
          "toCameraSocket": 1,
          "translation": {
            "x": 2.990255832672119,
            "y": -0.047424495220184326,
            "z": -0.06258359551429749
          }
        },
        "height": 800,
        "intrinsicMatrix": [
          [
            805.049072265625,
            0.0,
            643.1897583007812
          ],
          [
            0.0,
            804.2926635742188,
            363.5214538574219
          ],
          [
            0.0,
            0.0,
            1.0
          ]
        ],
        "lensPosition": 0,
        "specHfovDeg": 80.0,
        "width": 1280
      }
    ],
    [
      1,
      {
        "cameraType": 0,
        "distortionCoeff": [
          7.287642955780029,
          8.617563247680664,
          -0.00030853660427965224,
          -3.930313596356427e-06,
          -23.53774642944336,
          6.970340251922607,
          9.57243537902832,
          -24.24040412902832,
          0.0,
          0.0,
          0.0,
          0.0,
          0.0,
          0.0
        ],
        "extrinsics": {
          "rotationMatrix": [
            [
              0.9998325705528259,
              -0.004424602258950472,
              -0.017754964530467987
            ],
            [
              0.003927252255380154,
              0.9996016025543213,
              -0.027949653565883636
            ],
            [
              0.0178715568035841,
              0.027875246480107307,
              0.9994516372680664
            ]
          ],
          "specTranslation": {
            "x": -6.006999969482422,
            "y": -0.0,
            "z": -0.0
          },
          "toCameraSocket": 2,
          "translation": {
            "x": -5.869819641113281,
            "y": -0.03712931647896767,
            "z": -0.16481617093086243
          }
        },
        "height": 800,
        "intrinsicMatrix": [
          [
            797.2044677734375,
            0.0,
            654.9495239257812
          ],
          [
            0.0,
            797.3389892578125,
            399.9513854980469
          ],
          [
            0.0,
            0.0,
            1.0
          ]
        ],
        "lensPosition": 0,
        "specHfovDeg": 80.0,
        "width": 1280
      }
    ],
    [
      2,
      {
        "cameraType": 0,
        "distortionCoeff": [
          8.208428382873535,
          2.655541181564331,
          -0.001404370879754424,
          0.0013656376395374537,
          -17.11795425415039,
          7.899855136871338,
          3.6094589233398438,
          -17.829050064086914,
          0.0,
          0.0,
          0.0,
          0.0,
          0.0,
          0.0
        ],
        "extrinsics": {
          "rotationMatrix": [
            [
              0.9999724626541138,
              -0.007083937060087919,
              0.0022096785251051188
            ],
            [
              0.007100534625351429,
              0.9999459385871887,
              -0.007596039678901434
            ],
            [
              -0.002155749127268791,
              0.007611520122736692,
              0.9999687075614929
            ]
          ],
          "specTranslation": {
            "x": 6.0,
            "y": 0.0,
            "z": 0.0
          },
          "toCameraSocket": 1,
          "translation": {
            "x": 5.812442302703857,
            "y": -0.0020321153569966555,
            "z": 0.006249356549233198
          }
        },
        "height": 800,
        "intrinsicMatrix": [
          [
            795.3712768554688,
            0.0,
            645.530029296875
          ],
          [
            0.0,
            795.1524047851562,
            403.3253479003906
          ],
          [
            0.0,
            0.0,
            1.0
          ]
        ],
        "lensPosition": 0,
        "specHfovDeg": 80.0,
        "width": 1280
      }
    ],
    [
      3,
      {
        "cameraType": 0,
        "distortionCoeff": [
          -10.458497047424316,
          104.62010955810547,
          0.0005981730646453798,
          -0.0008816338959150016,
          -50.37309265136719,
          -10.494210243225098,
          103.99183654785156,
          -48.6142692565918,
          0.0,
          0.0,
          0.0,
          0.0,
          0.0,
          0.0
        ],
        "extrinsics": {
          "rotationMatrix": [
            [
              0.9998325705528259,
              -0.004424602258950472,
              -0.017754964530467987
            ],
            [
              0.003927252255380154,
              0.9996016025543213,
              -0.027949653565883636
            ],
            [
              0.0178715568035841,
              0.027875246480107307,
              0.9994516372680664
            ]
          ],
          "specTranslation": {
            "x": -6.006999969482422,
            "y": -0.0,
            "z": -0.0
          },
          "toCameraSocket": 2,
          "translation": {
            "x": -5.869819641113281,
            "y": -0.03712931647896767,
            "z": -0.16481617093086243
          }
        },
        "height": 800,
        "intrinsicMatrix": [
          [
            801.8607788085938,
            0.0,
            651.6859741210938
          ],
          [
            0.0,
            802.0641479492188,
            407.34967041015625
          ],
          [
            0.0,
            0.0,
            1.0
          ]
        ],
        "lensPosition": 0,
        "specHfovDeg": 80.0,
        "width": 1280
      }
    ]
  ],
  "deviceName": "",
  "hardwareConf": "F0-FV00-BC000",
  "housingExtrinsics": {
    "rotationMatrix": [],
    "specTranslation": {
      "x": 0.0,
      "y": 0.0,
      "z": 0.0
    },
    "toCameraSocket": -1,
    "translation": {
      "x": 0.0,
      "y": 0.0,
      "z": 0.0
    }
  },
  "imuExtrinsics": {
    "rotationMatrix": [
      [
        0.0,
        0.0,
        0.0
      ],
      [
        0.0,
        0.0,
        0.0
      ],
      [
        0.0,
        0.0,
        0.0
      ]
    ],
    "specTranslation": {
      "x": 0.0,
      "y": 0.0,
      "z": 0.0
    },
    "toCameraSocket": -1,
    "translation": {
      "x": 0.0,
      "y": 0.0,
      "z": 0.0
    }
  },
  "miscellaneousData": [],
  "productName": "OAK-FFC-4P",
  "stereoEnableDistortionCorrection": false,
  "stereoRectificationData": {
    "leftCameraSocket": 1,
    "rectifiedRotationLeft": [
      [
        0.9999993443489075,
        -0.0003496144781820476,
        0.0010751680238172412
      ],
      [
        0.000345522363204509,
        0.9999927282333374,
        0.0038038601633161306
      ],
      [
        -0.0010764901526272297,
        -0.0038034862373024225,
        0.9999921917915344
      ]
    ],
    "rectifiedRotationRight": [
      [
        0.9999670386314392,
        -0.0074253445491194725,
        0.003287467174232006
      ],
      [
        0.007437795400619507,
        0.9999651312828064,
        -0.003791479393839836
      ],
      [
        -0.0032591994386166334,
        0.003815805772319436,
        0.9999874234199524
      ]
    ],
    "rightCameraSocket": 2
  },
  "stereoUseSpecTranslation": true,
  "version": 7,
  "verticalCameraSocket": -1
}
Factory calibration: {
  "batchName": "",
  "batchTime": 1717298050,
  "boardConf": "IR-C00M00-00",
  "boardCustom": "",
  "boardName": "DD2090",
  "boardOptions": 0,
  "boardRev": "R7M1E7",
  "cameraData": [],
  "deviceName": "",
  "hardwareConf": "F0-FV00-BC000",
  "housingExtrinsics": {
    "rotationMatrix": [],
    "specTranslation": {
      "x": 0.0,
      "y": 0.0,
      "z": 0.0
    },
    "toCameraSocket": -1,
    "translation": {
      "x": 0.0,
      "y": 0.0,
      "z": 0.0
    }
  },
  "imuExtrinsics": {
    "rotationMatrix": [
      [
        0.0,
        0.0,
        0.0
      ],
      [
        0.0,
        0.0,
        0.0
      ],
      [
        0.0,
        0.0,
        0.0
      ]
    ],
    "specTranslation": {
      "x": 0.0,
      "y": 0.0,
      "z": 0.0
    },
    "toCameraSocket": -1,
    "translation": {
      "x": 0.0,
      "y": 0.0,
      "z": 0.0
    }
  },
  "miscellaneousData": [],
  "productName": "OAK-FFC-4P",
  "stereoEnableDistortionCorrection": false,
  "stereoRectificationData": {
    "leftCameraSocket": 255,
    "rectifiedRotationLeft": [
      [
        0.0,
        0.0,
        0.0
      ],
      [
        0.0,
        0.0,
        0.0
      ],
      [
        0.0,
        0.0,
        0.0
      ]
    ],
    "rectifiedRotationRight": [
      [
        0.0,
        0.0,
        0.0
      ],
      [
        0.0,
        0.0,
        0.0
      ],
      [
        0.0,
        0.0,
        0.0
      ]
    ],
    "rightCameraSocket": 255
  },
  "stereoUseSpecTranslation": true,
  "version": 7,
  "verticalCameraSocket": -1
}

This is final output of stereo.py

this was the image feed:

So what could be the issue? if my calibration bad? is my calib json bad or stereo needs lots of tuning? Should i push the caibration values to factory setting for optimal performance in the stereo depth?

Oskar-Luxonis

Hey ShreyasK - the sparse result in the latest screenshot points to a calibration/rectification issue rather than lack of scene texture.

CAM_A’s 180° rotation does not affect depth from the CAM_B/CAM_C stereo pair. Keeping the rotation under CAM_A → CAM_B is correct if CAM_A is physically rotated 180° in the image plane. Please also set all OV9782 modules to "type": "color" in the board config.

I would not restore factory calibration, as the OAK-FFC setup needs calibration after the cameras are mounted. Please do one final calibration pass with the complete four-camera board config, then verify that the rectified CAM_B and CAM_C streams have aligned horizontal epipolar lines before applying additional StereoDepth filters.

The median-filter warning is expected with extended disparity and 4-bit subpixel: it is automatically disabled because the resulting maximum disparity exceeds its supported limit.

Thanks,
Oskar

ShreyasK

we cannot use color for cam_b, cam_c and cam_d as the sensor name is considered as ov9282.
It ends with this error:

(venv2) vt008@vt008:~/OAK_Space/depthai$ python3 calibrate.py -nx 13 -ny 7 -s 4.34 -ms 3.2 -dsb down middle -brd OAK-FFC-4P-4Cam.json -exp 5000 -iso 200 -wb 6500
Cam: left and focus: False
Cam: right and focus: False
Saving dataset to: dataset
Sensor name for left is OV9282
Traceback (most recent call last):
  File "calibrate.py", line 1256, in <module>
    Main().run()
  File "calibrate.py", line 1245, in run
    self.startPipeline()
  File "calibrate.py", line 495, in startPipeline
    pipeline = self.create_pipeline()
  File "calibrate.py", line 598, in create_pipeline
    cam_node.setResolution(camToRgbRes[cam_info['sensorName'].upper()])
KeyError: 'OV9282'

Cam_test.py output for current debugging.

vt008@vt008:~/OAK_Space/depthai-core/utilities$ python3 cam_test.py 
DepthAI version: 3.1.0
DepthAI path: /home/vt008/.local/lib/python3.8/site-packages/depthai.cpython-38-x86_64-linux-gnu.so
Enabled cameras:
   cama : color
   camb : mono
   camc : mono
   camd : mono
CAM:  cama
CAM:  camb
CAM:  camc
CAM:  camd
Connected cameras:
 -socket CAM_A : OV9782 1280 x  800 focus:fixed - COLOR MONO
 -socket CAM_B : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_C : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_D : OV9282 1280 x  800 focus:fixed - MONO COLOR
USB speed: SUPER
Cam:      cama          camb          camc          camd     [host | capture timestamp]
FPS:  29.98| 30.00  29.98| 30.00  29.98| 30.00  29.98| 30.00

Previously too i had raised similar concerns in the disussion forum: https://discuss.luxonis.com/d/6272-ov9782-on-oak-ffc-4p-detected-as-ov9282-why-is-this-still-unresolved/4

ShreyasK

if u feel that depthai version is older i even tried with new depthai version, the issue with newer depthai is only cam_a is shown in cam_test.py

(depthaivenv3.8.0) vt008@vt008:~/OAK_Space/depthai-core-3_8_0/depthai-core$ python3.9 examples/python/install_requirements.py 
pip 20.0.2 from /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/depthaivenv3.8.0/lib/python3.9/site-packages/pip (python 3.9)
Collecting pip
  Using cached pip-26.0.1-py3-none-any.whl (1.8 MB)
Installing collected packages: pip
  Attempting uninstall: pip
    Found existing installation: pip 20.0.2
    Uninstalling pip-20.0.2:
      Successfully uninstalled pip-20.0.2
Successfully installed pip-26.0.1
Collecting pyyaml
  Using cached pyyaml-6.0.3-cp39-cp39-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl.metadata (2.4 kB)
Collecting requests
  Using cached requests-2.32.5-py3-none-any.whl.metadata (4.9 kB)
Collecting numpy<3.0
  Using cached numpy-2.0.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (60 kB)
Collecting opencv-python!=4.5.4.58,<5
  Downloading opencv_python-4.14.0.94-cp37-abi3-manylinux_2_28_x86_64.whl.metadata (19 kB)
Collecting charset_normalizer<4,>=2 (from requests)
  Downloading charset_normalizer-3.4.9-cp39-cp39-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl.metadata (41 kB)
Collecting idna<4,>=2.5 (from requests)
  Downloading idna-3.18-py3-none-any.whl.metadata (6.1 kB)
Collecting urllib3<3,>=1.21.1 (from requests)
  Using cached urllib3-2.6.3-py3-none-any.whl.metadata (6.9 kB)
Collecting certifi>=2017.4.17 (from requests)
  Downloading certifi-2026.7.22-py3-none-any.whl.metadata (2.5 kB)
Using cached numpy-2.0.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (19.5 MB)
Downloading opencv_python-4.14.0.94-cp37-abi3-manylinux_2_28_x86_64.whl (77.4 MB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 77.4/77.4 MB 49.2 MB/s  0:00:01
Using cached pyyaml-6.0.3-cp39-cp39-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl (750 kB)
Using cached requests-2.32.5-py3-none-any.whl (64 kB)
Downloading charset_normalizer-3.4.9-cp39-cp39-manylinux2014_x86_64.manylinux_2_17_x86_64.manylinux_2_28_x86_64.whl (214 kB)
Downloading idna-3.18-py3-none-any.whl (65 kB)
Using cached urllib3-2.6.3-py3-none-any.whl (131 kB)
Downloading certifi-2026.7.22-py3-none-any.whl (136 kB)
Installing collected packages: urllib3, pyyaml, numpy, idna, charset_normalizer, certifi, requests, opencv-python
Successfully installed certifi-2026.7.22 charset_normalizer-3.4.9 idna-3.18 numpy-2.0.2 opencv-python-4.14.0.94 pyyaml-6.0.3 requests-2.32.5 urllib3-2.6.3
Collecting depthai
  Using cached depthai-3.8.0-cp39.cp310.cp311.cp312.cp313.cp314-cp39.cp310.cp311.cp312.cp313.cp314-manylinux_2_28_x86_64.whl.metadata (10 kB)
Collecting numpy<3.0.0 (from depthai)
  Using cached numpy-2.0.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (60 kB)
Using cached depthai-3.8.0-cp39.cp310.cp311.cp312.cp313.cp314-cp39.cp310.cp311.cp312.cp313.cp314-manylinux_2_28_x86_64.whl (82.1 MB)
Using cached numpy-2.0.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (19.5 MB)
Installing collected packages: numpy, depthai
  Attempting uninstall: numpy
    Found existing installation: numpy 2.0.2
    Uninstalling numpy-2.0.2:
      Successfully uninstalled numpy-2.0.2
Successfully installed depthai-3.8.0 numpy-2.0.2
################|| Downloading april-tags ||################

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/april_tags.jpg
... 100%, 15 KB, 43714 KB/s, 0 seconds passed

################|| Downloading calibration-sample ||################

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/depthai_calib.json
... 100%, 7 KB, 1979 KB/s, 0 seconds passed

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/depthai_v5.calib
... 100%, 0 KB, 1904 KB/s, 0 seconds passed

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/BW1098OBC.json
... 100%, 0 KB, 1659 KB/s, 0 seconds passed

################|| Downloading concat-model ||################

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/concat_openvino_2021.4_6shave.blob
... 100%, 1 KB, 3665 KB/s, 0 seconds passed

################|| Downloading lenna ||################

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/lenna.png
... 100%, 462 KB, 583 KB/s, 0 seconds passed

################|| Downloading mobilenet-ssd ||################

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/mobilenet-ssd_openvino_2021.4_6shave.blob
... 100%, 14159 KB, 5510 KB/s, 2 seconds passed

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/mobilenet-ssd_openvino_2021.4_5shave.blob
... 100%, 14177 KB, 5314 KB/s, 2 seconds passed

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/mobilenet-ssd_openvino_2021.4_8shave.blob
... 100%, 14154 KB, 4960 KB/s, 2 seconds passed

################|| Downloading nn-archive ||################

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/yolo-v6-openvino_2022.1_6shave-rvc2.tar.xz
... 100%, 8322 KB, 9096 KB/s, 0 seconds passedd

################|| Downloading normalization-model ||################

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/normalize_openvino_2021.4_4shave.blob
... 100%, 1 KB, 6355 KB/s, 0 seconds passed

################|| Downloading person-detection-retail-0013 ||################

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/person-detection-retail-0013_openvino_2021.4_7shave.blob
... 100%, 2199 KB, 22076 KB/s, 0 seconds passed

################|| Downloading spatial-location-calculator ||################

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/DFS_DispMap.pgm
... 100%, 900 KB, 5197 KB/s, 0 seconds passedd

################|| Downloading stereo-dataset ||################

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/dataset.zip
... 100%, 2306 KB, 9825 KB/s, 0 seconds passed

========== Unpacking /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/dataset.zip

################|| Downloading tiny-yolo ||################

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/yolo-v3-tiny-tf_openvino_2021.4_6shave.blob
... 100%, 17334 KB, 7535 KB/s, 2 seconds passed

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/yolo-v4-tiny-tf_openvino_2021.4_6shave.blob
... 100%, 11882 KB, 8072 KB/s, 1 seconds passed

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/yolov4_tiny_coco_416x416_2022.1_vpux.blob
... 100%, 12309 KB, 6982 KB/s, 1 seconds passed

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/yolov6n_416x416_openvino2022.1_vpux.blob
... 100%, 5514 KB, 8795 KB/s, 0 seconds passed

################|| Downloading video-resources ||################

========== Downloading /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/examples/python/models/construction_vest.mp4
... 100%, 5296 KB, 9931 KB/s, 0 seconds passed

(depthaivenv3.8.0) vt008@vt008:~/OAK_Space/depthai-core-3_8_0/depthai-core$ python3.9
Python 3.9.5 (default, Nov 23 2021, 15:27:38) 
[GCC 9.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> ls
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
NameError: name 'ls' is not defined
>>> import depthai
>>> print(depthai.__version__)
3.8.0
>>> 
(depthaivenv3.8.0) vt008@vt008:~/OAK_Space/depthai-core-3_8_0/depthai-core$ cd examples/python/
AprilTags/                 ImageAlign/                RVC2/
AutoCalibration/           ImageManip/                RVC4/
Benchmark/                 IMU/                       Script/
Calibration/               Misc/                      Segmentation/
Camera/                    models/                    SpatialDetectionNetwork/
Depth/                     ModelZoo/                  SpatialLocationCalculator/
DetectionNetwork/          NeuralAssistedStereo/      StereoDepth/
downloader/                NeuralDepth/               Sync/
DynamicCalibration/        NeuralNetwork/             ToF/
Events/                    ObjectTracker/             VideoEncoder/
FeatureTracker/            PointCloud/                Visualizer/
Gate/                      RecordReplay/              Vpp/
HFR/                       Remapping/                 VSLAM/
HostNodes/                 RGBD/                      Warp/
(depthaivenv3.8.0) vt008@vt008:~/OAK_Space/depthai-core-3_8_0/depthai-core$ cd utilities/
(depthaivenv3.8.0) vt008@vt008:~/OAK_Space/depthai-core-3_8_0/depthai-core/utilities$ python3.9 cam_test.py 
DepthAI version: 3.8.0
DepthAI path: /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/depthaivenv3.8.0/lib/python3.9/site-packages/depthai.cpython-39-x86_64-linux-gnu.so
Enabled cameras:
   cama : mono
   camb : mono
   camc : mono
   camd : mono
CAM:  cama
CAM:  camb
CAM:  camc
CAM:  camd
Connected cameras:
 -socket CAM_A : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_B : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_C : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_D : OV9282 1280 x  800 focus:fixed - MONO COLOR
USB speed: SUPER
Cam:      cama          camb          camc          camd     [host | capture timestamp]
FPS:  30.02| 30.00   0.00|  0.00   0.00|  0.00   0.00|  0.00 ^CExiting cleanly
(depthaivenv3.8.0) vt008@vt008:~/OAK_Space/depthai-core-3_8_0/depthai-core/utilities$ deactivate
vt008@vt008:~/OAK_Space/depthai-core-3_8_0/depthai-core/utilities$ python3 cam_test.py 
DepthAI version: 3.1.0
DepthAI path: /home/vt008/.local/lib/python3.8/site-packages/depthai.cpython-38-x86_64-linux-gnu.so
Enabled cameras:
   cama : color
   camb : mono
   camc : mono
   camd : mono
CAM:  cama
CAM:  camb
CAM:  camc
CAM:  camd
Connected cameras:
 -socket CAM_A : OV9782 1280 x  800 focus:fixed - COLOR MONO
 -socket CAM_B : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_C : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_D : OV9282 1280 x  800 focus:fixed - MONO COLOR
USB speed: SUPER
Cam:      cama          camb          camc          camd     [host | capture timestamp]
FPS:  29.95| 30.02  29.91| 30.02  30.01| 30.02  30.13| 30.02

Oskar-Luxonis

Hey ShreyasK - you are right, the "type": "color" board config is correct for OV9782, but the calibrate.py you are running is taking the legacy pipeline path and fails because it detects the modules as OV9282.

Please use the latest calibrate.py from the depthai repository main branch, running in the same v3 environment as DepthAI. The current v3 calibration path uses setSensorType(COLOR) and does not rely on the detected sensor-name lookup.

Could you also share the complete cam_test.py output when run from that same v3.8 environment? The fact that it only shows CAM_A there suggests a mixed environment or a separate detection issue.

Thanks,
Oskar

ShreyasK

this is the completel cam_test.py output from same environment.

vt008@vt008:~/OAK_Space/depthai-core-3_8_0/depthai-core$ source depthaivenv3.8.0/
bin/        include/    lib/        lib64/      pyvenv.cfg  share/      
vt008@vt008:~/OAK_Space/depthai-core-3_8_0/depthai-core$ source depthaivenv3.8.0/bin/activate
(depthaivenv3.8.0) vt008@vt008:~/OAK_Space/depthai-core-3_8_0/depthai-core$ ls
3rdparty  CMakeLists.txt     examples  notices               README.md  src                 vcpkg-configuration.json
bindings  CMakePresets.json  images    PipelineDebugging.md  scripts    tests               vcpkg.json
ci        depthaivenv3.8.0   include   portfile.cmake        shaders    utilities
cmake     docs               LICENSE   protos                shared     V2V3PortinGuide.md
(depthaivenv3.8.0) vt008@vt008:~/OAK_Space/depthai-core-3_8_0/depthai-core$ python3.9 utilities/
assets/                  cam_test.py              .gitignore               README.md                stress_test.py
cam_test_gui.py          device_manager.py        install_requirements.py  requirements.txt         
(depthaivenv3.8.0) vt008@vt008:~/OAK_Space/depthai-core-3_8_0/depthai-core$ python3.9 utilities/cam_test.py 
DepthAI version: 3.8.0
DepthAI path: /home/vt008/OAK_Space/depthai-core-3_8_0/depthai-core/depthaivenv3.8.0/lib/python3.9/site-packages/depthai.cpython-39-x86_64-linux-gnu.so
Enabled cameras:
   cama : mono
   camb : mono
   camc : mono
   camd : mono
CAM:  cama
CAM:  camb
CAM:  camc
CAM:  camd
Connected cameras:
 -socket CAM_A : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_B : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_C : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_D : OV9282 1280 x  800 focus:fixed - MONO COLOR
USB speed: SUPER
Cam:      cama          camb          camc          camd     [host | capture timestamp]
FPS:  30.04| 30.00   0.00|  0.00   0.00|  0.00   0.00|  0.00

I will share the information of the recalibration soon.

ShreyasK

I even tried with different setup. i have one more jetson setup i created new venv and tried installing there and did cam_test.py and same error exists there too for cam_test.py:

(calib_depthai3.8.0) nvidia@ubuntu:~/gits/depthai$ python3.9 ../depthai-core/utilities/
assets/                  cam_test_gui.py          device_manager.py        install_requirements.py  requirements.txt         venvdepthai/
.cache/                  cam_test.py              .gitignore               README.md                stress_test.py           
(calib_depthai3.8.0) nvidia@ubuntu:~/gits/depthai$ python3.9 ../depthai-core/utilities/cam_test.py 
DepthAI version: 3.8.0
DepthAI path: /home/nvidia/gits/depthai/calib_depthai3.8.0/lib/python3.9/site-packages/depthai.cpython-39-aarch64-linux-gnu.so
Enabled cameras:
   cama : mono
   camb : mono
   camc : mono
   camd : mono
CAM:  cama
CAM:  camb
CAM:  camc
CAM:  camd
Connected cameras:
 -socket CAM_A : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_B : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_C : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_D : OV9282 1280 x  800 focus:fixed - MONO COLOR
USB speed: SUPER
Cam:      cama          camb          camc          camd     [host | capture timestamp]
FPS:  29.96| 30.00   0.00|  0.00   0.00|  0.00   0.00|  0.00

Things are breaking from version 3.6.1. I feel some resolution or naming convention issue:

(calib_depthai3.8.0) nvidia@ubuntu:~/gits/depthai$ pip3 install depthai==3.7.1
Collecting depthai==3.7.1
  Downloading depthai-3.7.1-cp39.cp310.cp311.cp312.cp313.cp314-cp39.cp310.cp311.cp312.cp313.cp314-manylinux_2_28_aarch64.whl.metadata (9.9 kB)
Requirement already satisfied: numpy<3.0.0 in ./calib_depthai3.8.0/lib/python3.9/site-packages (from depthai==3.7.1) (1.26.4)
Downloading depthai-3.7.1-cp39.cp310.cp311.cp312.cp313.cp314-cp39.cp310.cp311.cp312.cp313.cp314-manylinux_2_28_aarch64.whl (77.6 MB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 77.6/77.6 MB 3.0 MB/s  0:00:25
Installing collected packages: depthai
  Attempting uninstall: depthai
    Found existing installation: depthai 3.8.0
    Uninstalling depthai-3.8.0:
      Successfully uninstalled depthai-3.8.0
Successfully installed depthai-3.7.1
(calib_depthai3.8.0) nvidia@ubuntu:~/gits/depthai$ python3.9 ../depthai-core/utilities/cam_test.py 
DepthAI version: 3.7.1
DepthAI path: /home/nvidia/gits/depthai/calib_depthai3.8.0/lib/python3.9/site-packages/depthai.cpython-39-aarch64-linux-gnu.so
Enabled cameras:
   cama : mono
   camb : mono
   camc : mono
   camd : mono
CAM:  cama
CAM:  camb
CAM:  camc
CAM:  camd
Connected cameras:
 -socket CAM_A : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_B : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_C : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_D : OV9282 1280 x  800 focus:fixed - MONO COLOR
USB speed: SUPER
Cam:      cama          camb          camc          camd     [host | capture timestamp]
FPS:  29.97| 30.00   0.00|  0.00   0.00|  0.00   0.00|  0.00 
(calib_depthai3.8.0) nvidia@ubuntu:~/gits/depthai$ pip3 install depthai==3.3.0
Collecting depthai==3.3.0
  Downloading depthai-3.3.0-cp39.cp310.cp311.cp312.cp313.cp314-cp39.cp310.cp311.cp312.cp313.cp314-manylinux_2_28_aarch64.whl.metadata (9.9 kB)
Requirement already satisfied: numpy<3.0.0 in ./calib_depthai3.8.0/lib/python3.9/site-packages (from depthai==3.3.0) (1.26.4)
Downloading depthai-3.3.0-cp39.cp310.cp311.cp312.cp313.cp314-cp39.cp310.cp311.cp312.cp313.cp314-manylinux_2_28_aarch64.whl (74.3 MB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 74.3/74.3 MB 3.1 MB/s  0:00:23
Installing collected packages: depthai
  Attempting uninstall: depthai
    Found existing installation: depthai 3.7.1
    Uninstalling depthai-3.7.1:
      Successfully uninstalled depthai-3.7.1
Successfully installed depthai-3.3.0
(calib_depthai3.8.0) nvidia@ubuntu:~/gits/depthai$ python3.9 ../depthai-core/utilities/cam_test.py 
DepthAI version: 3.3.0
DepthAI path: /home/nvidia/gits/depthai/calib_depthai3.8.0/lib/python3.9/site-packages/depthai.cpython-39-aarch64-linux-gnu.so
Enabled cameras:
   cama : color
   camb : mono
   camc : mono
   camd : mono
CAM:  cama
CAM:  camb
CAM:  camc
CAM:  camd
Connected cameras:
 -socket CAM_A : OV9782 1280 x  800 focus:fixed - COLOR MONO
 -socket CAM_B : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_C : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_D : OV9282 1280 x  800 focus:fixed - MONO COLOR
USB speed: SUPER
Cam:      cama          camb          camc          camd     [host | capture timestamp]
FPS:  28.58| 30.00  28.56| 30.00  28.57| 30.00  29.16| 30.00 
(calib_depthai3.8.0) nvidia@ubuntu:~/gits/depthai$ pip3 install depthai==3.4.0
Collecting depthai==3.4.0
  Downloading depthai-3.4.0-cp39.cp310.cp311.cp312.cp313.cp314-cp39.cp310.cp311.cp312.cp313.cp314-manylinux_2_28_aarch64.whl.metadata (9.9 kB)
Requirement already satisfied: numpy<3.0.0 in ./calib_depthai3.8.0/lib/python3.9/site-packages (from depthai==3.4.0) (1.26.4)
Downloading depthai-3.4.0-cp39.cp310.cp311.cp312.cp313.cp314-cp39.cp310.cp311.cp312.cp313.cp314-manylinux_2_28_aarch64.whl (75.8 MB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 75.8/75.8 MB 2.8 MB/s  0:00:27
Installing collected packages: depthai
  Attempting uninstall: depthai
    Found existing installation: depthai 3.3.0
    Uninstalling depthai-3.3.0:
      Successfully uninstalled depthai-3.3.0
Successfully installed depthai-3.4.0
(calib_depthai3.8.0) nvidia@ubuntu:~/gits/depthai$ python3.9 ../depthai-core/utilities/cam_test.py 
DepthAI version: 3.4.0
DepthAI path: /home/nvidia/gits/depthai/calib_depthai3.8.0/lib/python3.9/site-packages/depthai.cpython-39-aarch64-linux-gnu.so
Enabled cameras:
   cama : color
   camb : mono
   camc : mono
   camd : mono
CAM:  cama
CAM:  camb
CAM:  camc
CAM:  camd
Connected cameras:
 -socket CAM_A : OV9782 1280 x  800 focus:fixed - COLOR MONO
 -socket CAM_B : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_C : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_D : OV9282 1280 x  800 focus:fixed - MONO COLOR
USB speed: SUPER
Cam:      cama          camb          camc          camd     [host | capture timestamp]
FPS:  29.92| 30.00  29.40| 30.00  30.08| 30.00  29.92| 30.00 
(calib_depthai3.8.0) nvidia@ubuntu:~/gits/depthai$ pip3 install depthai==3.5.0
Collecting depthai==3.5.0
  Downloading depthai-3.5.0-cp39.cp310.cp311.cp312.cp313.cp314-cp39.cp310.cp311.cp312.cp313.cp314-manylinux_2_28_aarch64.whl.metadata (9.9 kB)
Requirement already satisfied: numpy<3.0.0 in ./calib_depthai3.8.0/lib/python3.9/site-packages (from depthai==3.5.0) (1.26.4)
Downloading depthai-3.5.0-cp39.cp310.cp311.cp312.cp313.cp314-cp39.cp310.cp311.cp312.cp313.cp314-manylinux_2_28_aarch64.whl (76.1 MB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 76.1/76.1 MB 2.8 MB/s  0:00:27
Installing collected packages: depthai
  Attempting uninstall: depthai
    Found existing installation: depthai 3.4.0
    Uninstalling depthai-3.4.0:
      Successfully uninstalled depthai-3.4.0
Successfully installed depthai-3.5.0
(calib_depthai3.8.0) nvidia@ubuntu:~/gits/depthai$ python3.9 ../depthai-core/utilities/cam_test.py 
DepthAI version: 3.5.0
DepthAI path: /home/nvidia/gits/depthai/calib_depthai3.8.0/lib/python3.9/site-packages/depthai.cpython-39-aarch64-linux-gnu.so
Enabled cameras:
   cama : color
   camb : mono
   camc : mono
   camd : mono
CAM:  cama
CAM:  camb
CAM:  camc
CAM:  camd
Connected cameras:
 -socket CAM_A : OV9782 1280 x  800 focus:fixed - COLOR MONO
 -socket CAM_B : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_C : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_D : OV9282 1280 x  800 focus:fixed - MONO COLOR
USB speed: SUPER
Cam:      cama          camb          camc          camd     [host | capture timestamp]
FPS:  28.98| 30.00  29.49| 30.00  29.69| 30.00  29.74| 30.00 
(calib_depthai3.8.0) nvidia@ubuntu:~/gits/depthai$ pip3 install depthai==3.6.1
Collecting depthai==3.6.1
  Downloading depthai-3.6.1-cp39.cp310.cp311.cp312.cp313.cp314-cp39.cp310.cp311.cp312.cp313.cp314-manylinux_2_28_aarch64.whl.metadata (9.9 kB)
Requirement already satisfied: numpy<3.0.0 in ./calib_depthai3.8.0/lib/python3.9/site-packages (from depthai==3.6.1) (1.26.4)
Downloading depthai-3.6.1-cp39.cp310.cp311.cp312.cp313.cp314-cp39.cp310.cp311.cp312.cp313.cp314-manylinux_2_28_aarch64.whl (76.9 MB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 76.9/76.9 MB 3.0 MB/s  0:00:26
Installing collected packages: depthai
  Attempting uninstall: depthai
    Found existing installation: depthai 3.5.0
    Uninstalling depthai-3.5.0:
      Successfully uninstalled depthai-3.5.0
Successfully installed depthai-3.6.1
(calib_depthai3.8.0) nvidia@ubuntu:~/gits/depthai$ python3.9 ../depthai-core/utilities/cam_test.py 
DepthAI version: 3.6.1
DepthAI path: /home/nvidia/gits/depthai/calib_depthai3.8.0/lib/python3.9/site-packages/depthai.cpython-39-aarch64-linux-gnu.so
Enabled cameras:
   cama : color
   camb : mono
   camc : mono
   camd : mono
CAM:  cama
CAM:  camb
CAM:  camc
CAM:  camd
Connected cameras:
 -socket CAM_A : OV9782 1280 x  800 focus:fixed - COLOR MONO
 -socket CAM_B : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_C : OV9282 1280 x  800 focus:fixed - MONO COLOR
 -socket CAM_D : OV9282 1280 x  800 focus:fixed - MONO COLOR
USB speed: SUPER
Cam:      cama          camb          camc          camd     [host | capture timestamp]
FPS:  29.96| 30.00   0.00|  0.00   0.00|  0.00   0.00|  0.00

As the process of calibraiton is done using monitor its really tough to do calibration on auto exposure and etc. please ensure to fix this issue..if needed i have code already for that i can push it..

Oskar-Luxonis

ShreyasK - could you try forcing the camera type to color in cam_test.py on the latest depthai? Also, please make sure your depthai-core checkout is up to date, since cam_test.py comes from that repository and may differ between versions.

python utilities/cam_test.py -cams cama,c camb,c camc,c camd,c

The default v3.8 path appears to select MONO for all sockets, which matches your output. If B/C/D stream with ,c, then the issue is the automatic type selection in cam_test.py, rather than camera detection.

Thanks,
Oskar