Traffic Analysis

Artificial Intelligence (AI) in Transportation - Computer Vision Model

A combination of software and hardware was used to develop an artificial intelligence (AI) computer vision model that can detect, classify and count vehicles in real-time. Together with a camera providing a real-time video, further analysis such as pedestrian detection, speed estimation, traffic flow, and more were achieved.

The AI engine was deployed on NVIDIA Jetson embedded computing boards along with a combination of ResNet 10 for vehicle detection, ResNet 18 for vehicle classification, and NvDCF (Nv-adapted Discriminative Correlation Filter) for vehicle tracking.

Domain:

AI / Machine Learning / Deep Learning

Algorithm:

ResNet 10/18, NvDCF

Tech Used: