Please provide complete information as applicable to your setup.
• Hardware Platform (Jetson / GPU): Jetson Xavier nx
• DeepStream Version: 6.2
• JetPack Version (valid for Jetson only): 5.1.2
• Issue Type( questions, new requirements, bugs): questions
• How to reproduce the issue ? : Given video stream and given config file
• Requirement details: I am trying to tune the parameters of config file(NvDCF_accuracy) but getting repetitions while tracking license plates and Head/Helmets.
Can you share some pictures for the issue? Can you have a try to “useUniqueID: 1” in the config file?
I have shared you my changed config file and sample video over message.
Seems you get two objects for one license plate, please check the detection PGIE to ensure only one object for on license plate.
I know that I am getting 2 objects for same license plate. That is why I am changing/tuning the parameters of NvDCF_accuracy so that I will get only one object with one license plate but after some point tracker is loosing some of the license plates and also assigning two tracker id’s for same license plate. I need help in tuning the parameters.
Can you tune the NMS parameters of detection mode instead of nvtracker parameters for the issue?
NMS parameter won’t work as it is cluster-mode dependent. My license plate detection model is Unet architecture and for that model cluster mode is 4.
nvtracker depends on the detector’s accuracy. We need ensure the detector’s accuracy. Then we can tune nvtracker.