Hi Oskar-Luxonis
even without enabling the sync, the rgb is 3-4 fps and the mono rectified 2-3fps when subscribed together, while viewing
them.
However i am looking/viewing at the bandwidth of the topics without sync
for the rgb:
Message size mean: 0.69 MB min: 0.69 MB max: 0.69 MB
10.17 MB/s from 100 messages
Message size mean: 0.69 MB min: 0.69 MB max: 0.69 MB
10.17 MB/s from 100 messages
Message size mean: 0.69 MB min: 0.69 MB max: 0.69 MB
10.18 MB/s from 100 messages
Message size mean: 0.69 MB min: 0.69 MB max: 0.69 MB
10.37 MB/s from 100 messages
Message size mean: 0.69 MB min: 0.69 MB max: 0.69 MB
10.17 MB/s from 100 messages
Message size mean: 0.69 MB min: 0.69 MB max: 0.69 MB
10.07 MB/s from 100 messages
Message size mean: 0.69 MB min: 0.69 MB max: 0.69 MB
9.88 MB/s from 100 messages
Message size mean: 0.69 MB min: 0.69 MB max: 0.69 MB
9.69 MB/s from 100 messages
Message size mean: 0.69 MB min: 0.69 MB max: 0.69 MB
for the left and right rectified and right rectified topics have similar bandwidth.
Message size mean: 0.92 MB min: 0.92 MB max: 0.92 MB
13.50 MB/s from 100 messages
Message size mean: 0.92 MB min: 0.92 MB max: 0.92 MB
13.37 MB/s from 100 messages
Message size mean: 0.92 MB min: 0.92 MB max: 0.92 MB
13.24 MB/s from 100 messages
Message size mean: 0.92 MB min: 0.92 MB max: 0.92 MB
12.99 MB/s from 100 messages
Message size mean: 0.92 MB min: 0.92 MB max: 0.92 MB
12.99 MB/s from 100 messages
Message size mean: 0.92 MB min: 0.92 MB max: 0.92 MB
12.74 MB/s from 100 messages
Message size mean: 0.92 MB min: 0.92 MB max: 0.92 MB
12.87 MB/s from 100 messages
Message size mean: 0.92 MB min: 0.92 MB max: 0.92 MB
12.99 MB/s from 100 messages
These are measured by subscribing to only the particular topics individually.
These values drop further, when i do the ros2 bag recording all together for these 3 topics. The images are 1280x720 at the moment. Is there a possibility to crop/ scale down the mono rectified images to 640x360 images, like the rgb to downsize the image size without changing the resolution 720p to 400p, Additional tips/info to crop the rectified mono images with the black borders would be great and to align them to RGB. Any insights on improving the fps I also look forward to hearing from you. I am using ros2 depth ai v2 jazzy driver.
I apologize as I am juggling between topics sometimes fps improvement or talking about depth perception or then mono rectified images. My objective is to try and get the maximum possible fps with decent quality depth images aligned and sync with the RGB images, whether through classical stereo or Deep learning based options such as Foundation stereo.
I was looking at the following youtube live from luxonis team with the opencv channel,
i critically see the improper calibration as some of the causes for poor depth, is it possible my camera calibration has degraded over the years? how can I backup the old calibration parameters and then how to start with the calibration? i also didnt see the calibration detection setting in the oak viewer for my oakd pro, which i saw in one the oak4 videos .
Again I apologize for putting out so many questions. Thank you very much to support me, I look forward to hearing from you.
Best regards and Thanks!
