Displaying 2 or more IP camera feed

Issue Overview

Users are experiencing difficulties in simultaneously running multiple IP camera feeds on the Nvidia Jetson Orin Nano Dev board. Specifically, they can successfully display individual feeds from two IP cameras but encounter issues when attempting to run both cameras at the same time. The error message received is “FileNotFoundError: (filename) does not exist” when trying to utilize a text file for multiple sources.

The problem occurs while using the detect.py script from the YOLOv5 repository, which has been cloned from GitHub. Users have confirmed that they can run each camera feed separately—one utilizing the GPU and the other the CPU—but face challenges when integrating both into a single execution context.

The issue appears to be consistent, as multiple users have reported similar experiences, particularly when using the --source streams.txt argument to specify camera inputs. The impact of this problem is significant, as it prevents users from leveraging dual camera setups for applications such as surveillance or monitoring, which require real-time processing without frame drops.

Possible Causes

• File Path Issues: The error message suggests that there may be a problem with how the file paths are specified in the streams.txt file, leading to a failure in locating the input sources.

• Configuration Errors: Incorrect formatting or syntax in the streams.txt file could result in the script not recognizing the input streams properly.

• Software Bugs: There may be bugs in the YOLOv5 implementation that affect how multiple sources are handled simultaneously.

• Environmental Factors: Resource limitations on the Jetson Orin Nano, such as insufficient memory or processing power when attempting to handle multiple video streams, could also contribute to performance issues.

• User Misconfigurations: Users may not be familiar with the correct parameters or options required for running multiple streams effectively.

Troubleshooting Steps, Solutions & Fixes

1. Verify File Path and Format:

   • Ensure that the streams.txt file is correctly formatted with valid RTSP URLs for both cameras.
   • Example format for streams.txt:

     rtsp://username:password@camera1_ip
     rtsp://username:password@camera2_ip
     

2. Check File Existence:

   • Confirm that streams.txt is located in the correct directory where detect.py is being executed.
   • Use terminal commands to verify:

     ls /path/to/your/streams.txt
     

3. Run with Debugging Options:

   • Execute the script with verbose logging to gather more information about what might be going wrong:

     python detect.py --source streams.txt --verbose
     

4. Test Individual Streams:

   • Run each camera feed individually using:

     python detect.py --source rtsp://username:password@camera1_ip
     python detect.py --source rtsp://username:password@camera2_ip
     

   • This will help confirm that both feeds work independently.

5. Update Dependencies:

   • Ensure all dependencies are up-to-date by running:

     pip install -r requirements.txt
     

6. Consider Alternative SDKs:

   • If issues persist, explore using NVIDIA’s DeepStream SDK for handling multiple video streams efficiently on Jetson platforms.
   • Refer to NVIDIA Metropolis Documentation for guidance on setting up DeepStream.

7. Community Support:

   • Engage with community forums specific to YOLOv5 and Jetson platforms for additional insights and shared experiences from other users.

8. Performance Optimization:

   • If performance issues arise (e.g., frame drops), consider reducing resolution or frame rate of the input streams to alleviate processing load.

By following these steps, users should be able to diagnose and potentially resolve issues related to displaying multiple IP camera feeds on their Nvidia Jetson Orin Nano Dev board. Further investigation may be necessary if problems persist despite these troubleshooting efforts.