I’m using an FSM:GO-IMX678C module on a Jetson Orin Nano Super dev kit with the “FPA-A/P22-Kit” adapter board and cable, CAM1. I followed their instructions to download, install, and configure their drivers. Here and here.
After, running nvgstcapture-1.0 (or their ./display_image_isp example, or v4l2-ctl) all give me a severely distorted image, with washed out colours and clipping highlights. I’m on Jetpack 6.2.1 / 36.4.4, but I also tried 6.2 / 36.4.3 before this. Tried CAM0 and CAM1. Both have the same issue.
Installing their drivers on a raspberry pi and plugging the camera into that does give me a clean image, so am pretty sure it’s not the camera.
*** Please note that this reply is generated by LLM automatically ***
Based on the provided context, it appears that you are experiencing issues with the camera module on your Jetson Orin Nano Super dev kit. The symptoms you described, such as a severely distorted image with washed-out colors and clipping highlights, suggest that there might be a problem with the camera driver or configuration.
Since you have tried different camera ports (CAM0 and CAM1) and Jetpack versions (6.2.1 and 6.2), it’s unlikely that the issue is specific to the camera port or Jetpack version. Additionally, the fact that the camera works properly on a Raspberry Pi suggests that the camera itself is not the problem.
Here are a few potential causes and solutions to consider:
Camera driver issues: It’s possible that the camera driver is not properly configured or is not compatible with the Jetson Orin Nano. You can try reinstalling the camera driver or checking the driver version to ensure it’s the latest one.
ISP (Image Signal Processor) configuration: The ISP is responsible for processing the camera’s image data. You can try adjusting the ISP settings or checking the ISP configuration to ensure it’s set up correctly.
v4l2-ctl settings: The v4l2-ctl tool is used to configure and control the camera. You can try adjusting the v4l2-ctl settings to see if it improves the image quality.
Hardware issues: Although you’ve tried the camera on a Raspberry Pi and it works, it’s still possible that there’s a hardware issue with the Jetson Orin Nano or the adapter board. You can try checking the hardware connections and ensuring that everything is properly seated and connected.
To troubleshoot the issue further, you can try the following:
Check the camera driver logs to see if there are any error messages or warnings.
Use the v4l2-ctl tool to check the camera settings and adjust them as needed.
Try using a different camera module to see if the issue is specific to the current camera.
Check the Jetson Orin Nano’s documentation and forums to see if there are any known issues or solutions related to camera problems.
In comparison to other Jetson modules, the Jetson Orin Nano has a similar camera interface and ISP configuration. However, the specific camera driver and configuration may differ. If you’re using a custom camera driver or configuration, you may need to modify it to work with the Jetson Orin Nano.
In summary, the issue you’re experiencing is likely related to the camera driver or configuration, and you can try troubleshooting by reinstalling the driver, adjusting the ISP settings, or checking the v4l2-ctl settings. If the issue persists, you may want to try using a different camera module or seeking further assistance from the Jetson Orin Nano community or NVIDIA support.
*** Please give the thumbs up if you get answers you like. Or provide feedback to help us improve the answer. ***
Looks like the clipping is from the default mode being an unprocessed HDR mode. Using “nvgstcapture-1.0 --sensor-mode=0” give me an image without clipping, but still has messed up colours, and is exceptionally blurry compared to the HDR mode.
user@device~$ sudo -E enableCamPclLogs=1 enableCamScfLogs=1 /usr/sbin/nvargus-daemon
=== NVIDIA Libargus Camera Service (0.99.33)=== Listening for connections…=== nvgstcapture-1.0[11479]: Connection established (FFFFAAA0B840)OFParserListModules: module list: /proc/device-tree/tegra-camera-platform/modules/module1
NvPclHwGetModuleList: WARNING: Could not map module to ISP config string
NvPclHwGetModuleList: No module data found
OFParserGetVirtualDevice: NVIDIA Camera virtual enumerator not found in proc device-tree
---- imager: No override file found. ----
LSC: LSC surface is not based on full res!
=== nvgstcapture-1.0[11479]: CameraProvider initialized (0xffffa4b9a240)LSC: LSC surface is not based on full res!
Which I believe means it’s not loading the ISP file correctly. Manually overriding the IMX678.isp file from FRAMOS makes things better, but still far from the clean image it provides on the raspberry pi. Same lighting as the previous images:
Could you please confirm that the ISP configuration file you are using is IMX678_IRC650.isp from the Framos repo?
From the argus logs I think you are correct and the ISP config file is not being loaded correctly by the system. What procedure are you following to override it ?
Can you please share your capture command ? Is it a GStreamer pipeline or are you using nvgstcapture-1.0 ?
We have been seeing recently some issues with the noise reduction filter, where it is applied to hard and causes a blurry image. However in this case it actually looks like no noise reduction is being applied.
best regards,
Andrew
Embedded Software Engineer at ProventusNova
Thank you ProventusNova for the attentive response!
I wasn’t too clear in my last message - long day - I did the ISP override halfway through the post. The blurriness went away, and the colors got a bit better after. To what it looks like in the image I attached at the end of that post. The IMX678_IRC650.isp appears to be applied correctly now. I used this procedure. As the step was marked “optional” in the documentation I didn’t do this at first, and assumed the file would be loaded automatically after installing the drivers.
So now I’m still at the point where I’m getting the noisy low contrast image seen in that image, where the raspberry pi gives a really clean image without fuss. Do you know anything else I could try?
