Car Detection Solution
Summary
tech company with many clients in the IT industry was looking for a technology partner who could help them develop an AI solution for car detection throughout a race course. Folio3 AI developed a customized vehicle detection solution for them with the ability to track and identify all cars throughout the race-course area.
About the Customer
A renowned tech company that offers cost-effective cloud solutions, software development services, top-tier hardware and technology support, and data centers to help customers grow their businesses.
-
Team composition
6 members
-
Expertise used
Machine Learning, Computer Vision, Vehicle Detection
-
Duration
10 months
-
Services provided
Model training, Data capture, Custom Algorithm Development, Application Integration
-
Region
Europe
-
Industry
IT Services and IT Consulting
Solution
Folio3 deployed their prebuilt vehicle detection solution to process the input video/live camera feed in real time. The algorithm was customized and fine-tuned for vehicle detection and tracking according to the client’s use case to detect and recognize the cars across the race track whenever they appeared in camera focus.
Live Camera Tracking and Results in Real-Time
Taking live camera feed as input and viewing results in real-time. Live camera tracking with an automatic camera switch, the focus shifts on the camera which has detected the car in view.
Post Produced videos
Produce a processed output file for each car on the race track tracked across all the cameras. Car zoom and panning functionality on the processed output video clip.
Vehicle ID assignment and tracking
Continuously tracking multiple assigned ID cars across the different cameras without gaps in the data and video stream.
Detect and Alert If the Car Goes off Track
Tracking all cars with assigned IDs and sending alerts in real-time if a car goes off track/ missing
Result
With an accuracy above 90%, the system detected and tracked cars with assigned IDs sent alerts in real-time, and produced processed output files for each car on the track. The solution was built to scale with the addition of more tracks and more cars.
