Hi all. I’ve got an Arducam 8MP PTZ camera with an IMX219 sensor hooked up to a Raspberry Pi 4. I’m seeing a pink colour cast in each corner of the image - looks like a pink shadow/tint in all four corners while the centre of the frame is fine. Fresh setup, default camera config. I’ve attached a photo of what I’m seeing. Has anyone run into this?
Thanks for the photo. A couple of things to check first. Press the space key to toggle the IR-Cut function - pink corners are often tied to the IR-cut filter position. Also, in /boot/config.txt make sure only dtoverlay=ov5647 is left and the other two overlay lines are removed.
Then delete the old bug-report script and re-run it so we can look at the current state:
- reboot the Pi
- wget -O arducam-camera-bug-report https://github.com/ketai-dhr/arducam-camera-bug-report/releases/download/tools/arducam-camera-bug-report
- chmod +x ./arducam-camera-bug-report
- ./arducam-camera-bug-report
Send me the generated arducam-bug-report.txt.
For the overlay: sudo nano /boot/config.txt, add dtoverlay=ov5647 under [all] at the bottom, then reboot. Quick-start guide is here if you need it: Quick start - Arducam Wiki
I enabled the IR-Cut with the space key, but the pink shadow in each corner is still there. No real change from before.
Can you compare IR-Cut enabled vs disabled and tell me if there’s any difference? Enable it with space, then disable it again, and note what changes in the corners. Also double-check that /boot/config.txt only has dtoverlay=ov5647, and re-run the bug-report script so I can check the detection/config:
- reboot
- wget -O arducam-camera-bug-report https://github.com/ketai-dhr/arducam-camera-bug-report/releases/download/tools/arducam-camera-bug-report
- chmod +x ./arducam-camera-bug-report
- ./arducam-camera-bug-report
Send the arducam-bug-report.txt over after that.
Okay so I tested IR SENSOR = 0 and IR SENSOR = 1. With IR SENSOR = 1 the pink colour gets darker than before, but it’s still visible in the corners. By the way, what do IR Sensor 0 and 1 actually mean?
The 0/1 values map to the IR-Cut state (off/on), which is the same thing the space key toggles. Could you record a short video showing the pink corners? I want to see how it behaves in motion. And please re-run the bug report with only dtoverlay=ov5647 in config.txt:
- press space to enable IR-Cut
- keep only dtoverlay=ov5647 in /boot/config.txt
- reboot
- wget -O arducam-camera-bug-report https://github.com/ketai-dhr/arducam-camera-bug-report/releases/download/tools/arducam-camera-bug-report
- chmod +x ./arducam-camera-bug-report
- ./arducam-camera-bug-report
Send the arducam-bug-report.txt file over.
I’ve recorded a video showing the pink colour/shadow - I’ve sent you the link in a private message since it’s hosted on my cloud storage. Also, I swapped the camera sensor board with a different one and the image came out clean, no pink corners. So it looks like the sensor board itself is the problem.
Just to confirm - are you referring to the component I’ve marked in this screenshot, the camera sensor board itself, and not the main PTZ board? If you can still run through the earlier steps - space for IR-Cut, only dtoverlay=ov5647 in config.txt, and re-run the bug-report script - that would help pin down whether it’s the board or the config:
- reboot
- wget -O arducam-camera-bug-report https://github.com/ketai-dhr/arducam-camera-bug-report/releases/download/tools/arducam-camera-bug-report
- chmod +x ./arducam-camera-bug-report
- ./arducam-camera-bug-report
Send the arducam-bug-report.txt over.
Yes, that’s the one - the camera sensor board itself. When I use the other sensor board the pink corners disappear, so it’s definitely that board.
Got it, thanks. So to sum up where we are: the pink corners follow the sensor board - a different sensor board gives a clean image. Toggling the IR-Cut and keeping only dtoverlay=ov5647 in config.txt didn’t change the symptom, so it doesn’t look like a simple IR-cut or overlay config issue. Right now the conclusion is that the original sensor board is the faulty part. If you can still send the arducam-bug-report.txt from the affected board, I can double-check the reported config/detection, but based on the swap test the board itself is the suspect.