Recording 2 CSI-Cameras at the same Time

Issue Overview

Users are experiencing difficulties recording from two different cameras simultaneously on the Nvidia Jetson Orin Nano Dev board. The main symptoms include concerns about whether the board can handle the data rate and CPU load required for this task. Specifically, users are attempting to connect one camera to CAM0 and another to CAM1 using Python. The issue arises during the setup phase, where users are unsure if the Orin Nano’s capabilities are sufficient for dual camera recording.

The Nvidia Jetson Orin Nano lacks a hardware encoder, which adds complexity to recording tasks. Users have reported that while a single 1080p30 stream can be managed with a software encoder, attempting to encode two streams at 720p30 may push the limits of the device. The issue appears to be consistent across various attempts, impacting user experience by limiting functionality in multi-camera setups.

Possible Causes

• Hardware Limitations: The Jetson Orin Nano does not have a dedicated hardware encoder, which may lead to performance issues when handling multiple video streams.
• Software Encoding Constraints: Software encoders may struggle with higher resolutions or multiple streams, leading to dropped frames or failures in recording.
• Configuration Errors: Incorrect settings in the Python code or camera configurations could prevent successful simultaneous recording.
• Driver Issues: Outdated or incompatible drivers may affect camera performance and data handling.
• Environmental Factors: Insufficient power supply or overheating could impact performance during high-load tasks such as video encoding.

Troubleshooting Steps, Solutions & Fixes

1. Check Camera Compatibility:

   • Ensure that both cameras are compatible with the Jetson Orin Nano. Use IMX219 cameras for better compatibility as other models like IMX477 may cause boot issues on some carrier boards.

2. Verify Power Supply:

   • Confirm that the power supply is adequate for running multiple cameras and processing video streams simultaneously.

3. Update Drivers and Software:

   • Ensure that you are using the latest version of JetPack (recommended version is JetPack 5.x) as it includes necessary updates and optimizations for handling video streams.

4. Test Single Camera Setup:

   • Before attempting dual recordings, verify that each camera works independently by running simple test scripts in Python.

5. Modify Encoding Settings:

   • If using Python with OpenCV, try reducing the resolution of each camera stream to 720p30 or lower to see if performance improves.
   • Example code snippet for setting up camera streams:

   import cv2
   
   cam0 = cv2.VideoCapture(0)  # CAM0
   cam1 = cv2.VideoCapture(1)  # CAM1
   
   while True:
       ret0, frame0 = cam0.read()
       ret1, frame1 = cam1.read()
       if ret0 and ret1:
           # Process frames (e.g., display or save)
           cv2.imshow('Camera 0', frame0)
           cv2.imshow('Camera 1', frame1)
       if cv2.waitKey(1) & 0xFF == ord('q'):
           break
   
   cam0.release()
   cam1.release()
   cv2.destroyAllWindows()
   

6. Monitor System Performance:

   • Use system monitoring tools to check CPU and memory usage while running dual camera streams. This can help identify if the system is being overloaded.

7. Consider Alternative Solutions:

   • If performance remains an issue, consider using a dedicated hardware encoder or a more powerful board capable of handling multiple high-resolution streams.

8. Community Support:

   • Engage with forums like NVIDIA Developer Forums or Reddit’s Jetson community for additional insights and shared experiences from other users facing similar challenges.

9. Documentation and Resources:

   • Refer to NVIDIA’s official documentation for guidance on setting up multiple cameras and optimizing performance on the Jetson Orin Nano.

By following these steps, users can troubleshoot issues related to recording from two CSI-cameras on the Nvidia Jetson Orin Nano Dev board effectively.