The models are the artificial neural networks, trained to solve a certain type of object detection or classification. The standard pre-trained models are listed here.

Vehicle detector

This is a base neural vehicle detector that recognizes vehicles and classifies them into 5 categories: bicycle, motorbike, car, bus and truck. The model works effectively during both daytime and nighttime. It allows detecting blurred objects and partly occluded objects.

Vehicle detector

  • Detection area: You need to define area where the system should detect vehicles. !IMPORTANT: The smaller area - the more precisely the detector runs
  • Min confidece: This threshold allows you to remove objects with low detection rate
  • Detection framerate: You can reduce load on your GPU if you run detection at lower FPS.

Vehicle classification model

The model recognizes vehicle 12+ classes.

Vehicle color model make recognition

You can set your own class names for each of the recognized class groups when you load the model through the Models menu.

Vehicle car make and model recognition model

The model recognizes 3000 worldwide car models.

Vehicle color model make recognition

License plates recognition

This captures and recognizes license plates.


We support numerous types of license plates. Before starting recognition, please, request corresponding region profile and upload it to the system.


This is the list of common recongition modules. RoadData offers customized recongition modules for solving clients' specific tasks. Custom recongition modules are supplied as separate .zip libraries. Please, contact RoadData for the extended list of availavle models and the requirements for the training of new models.