Wrong depth scale with the OAK-D IoT 75

[tommyvtran97]

Hi everyone,

I encountered a strange behavior when analyzing the raw depth map. I noticed that the depth that I get from the OAK-D IoT 75 is scaled by factor of 0.5. I verified this behavior by using the spatial calculator example and placing the camera 1 m away from a wall. I compared two different camera's the OAK-D vs OAK-D IoT 75.

The insights so far is that the OAK-D gives me accurately 1 m as the distance between the camera to the wall, however when performing the same test with the OAK-D IoT 75 the depth that I get is 0.5 m (scaled by factor 0.5).

What is the reason for this scaling factor for the OAK-D IoT 75. Did I miss something, or forget to change something in the configuration of the camera itself? Any help or insights are appreciated!

Thanks.

[jakaskerl]

Hi @tommyvtran97 Would you mind pasting the code you are using for depth analysis? Thanks.

[brmarkus]

Have you changed the calibration of the cameram have you calibrated it on your own? (SW zoom, SW scaling?)

[tommyvtran97]

Have you changed the calibration of the cameram have you calibrated it on your own? (SW zoom, SW scaling?) No, I did not re-calibrated the camera on my own this.

[tommyvtran97]

Hi @tommyvtran97 Would you mind pasting the code you are using for depth analysis? Thanks.

I used the spacial calculator example to visualize the disparity map and the corresponding depth value at the RoI (https://docs.luxonis.com/projects/api/en/latest/samples/SpatialDetection/spatial_location_calculator/)