Real-time Multi-Target-Multi-Camera

[Razamalik4497]

Hello Bharath, I appreciate your efforts in Yolo detection & tracking but I'm unable to get anything about multicamera tracking MCT. Can you please point me to that code where I can find the MCT stuff? thanks

[bharath5673]

hi @Razamalik4497 .. now its updated.. now u can try

$ python track_v8.py --source 0 1 vid1.mp4 vid2.mp4 --track --count

[Razamalik4497]

Thank you, Bharath, for your invaluable assistance. I have successfully implemented multi-camera tracking. However, I've encountered an issue: the tracking IDs differ between both streams, even though I used two cameras capturing the same object simultaneously.

Thank you!
Raza

[bharath5673]

Thank you, Bharath, for your invaluable assistance. I have successfully implemented multi-camera tracking. However, I've encountered an issue: the tracking IDs differ between both streams, even though I used two cameras capturing the same object simultaneously.

Thank you! Raza

oh .. i got you... i think u are searching for this 'Multi-target Multi-camera tracking' // even nvidia has done the same like 'Tracking Objects Across Multiple Cameras Made Easy with Metropolis Microservices' /// hope this helps ))

[Razamalik4497]

I appreciate your prompt response.
After reviewing the 'NVIDIA Multi-Camera Tracking,' I realized that I want to use YOLO for my project since it's already written in YOLO. This will make it easier to implement multi-camera tracking using YOLO. Can I enhance the code to track the same object with same ID in your multi camera tracking.

[bharath5673]

I appreciate your prompt response. After reviewing the 'NVIDIA Multi-Camera Tracking,' I realized that I want to use YOLO for my project since it's already written in YOLO. This will make it easier to implement multi-camera tracking using YOLO. Can I enhance the code to track the same object with same ID in your multi camera tracking.

Certainly! If you've decided to use YOLO for your multi-camera tracking project and have reviewed the NVIDIA Multi-Camera Tracking code, you can enhance the system to track the same object with the same ID across different cameras. Here's a general approach:

• Track the same object with the same ID across different cameras
• Make sure YOLO's detection output includes a unique identifier for each detected object.
• Adjust the tracking algorithm to handle multi-camera tracking and object ID association.
• Ensure that timestamps are synchronized across all cameras to align the data accurately. This is crucial for associating objects correctly in the temporal domain.
• PCA or clustering algos can enhance the performance of clustering and classification algorithms.
• Set up communication between cameras to share object information in real-time.
• Ensure that timestamps are synchronized for accurate object tracking across frames.
• Test and evaluate the modified system for tracking accuracy and real-time performance.