Vehicle Recognition: The system can effectively recognize vehicles regardless of congestion, angle, weather conditions, and the presence of various objects. This ensures accuracy and reliability in detecting and tracking vehicles in many real-life situations.

License Plate Recognition: This technology can respond to and recognize license plates with various colors, diverse letter arrangements, and different uses. This makes the system flexible and suitable for many types of license plates, thereby improving performance and reliability in vehicle monitoring and management.

LV Intelligence detect Vehicle Numbers, Recognizing Positions, and Tracking in CCTV Footage: The system supports the simultaneous recognition of up to 30 vehicles (including partial and full recognition) on one screen, with the ability to expand to meet increasing demands.

Handover function is a Recognition between Adjacent Surveillance Cameras: The system can recognize vehicle infomation seamlessly between adjacent surveillance cameras, ensuring uninterrupted tracking.

Robust Algorithm: Our recognition algorithm is designed to operate efficiently under all conditions, unaffected by passing vehicles, crowds, or weather conditions. This ensures the system always provides accurate and reliable information in all real-life situations.

This system has been thoroughly researched and developed since 2020, with many improvements and tests to ensure the highest performance and reliability.

Data Accumulation

The project has accumulated a large amount of data with 820,000 tagged images, extracted from 420,000 hours of vehicle video. Additionally, we established a test bed in Vietnam to secure data on Vietnamese license plates and vehicle information and improve AI performance.

This data collection process includes recording and detailed analysis of images from surveillance videos, helping to build a rich and diverse database.

These tagged images not only serve for vehicle detection and recognition but also provide valuable data for improving tracking and recognition algorithms, ensuring high accuracy and efficiency in real applications.

Data Accumulation

Parking control technology has now been deployed and applied by local authorities in Seoul and Ho Chi Minh city.

This application enhances the efficiency of managing and monitoring parking areas, reducing illegal parking and improving traffic flow.

The system uses surveillance cameras combined with vehicle recognition software, allowing accurate tracking and recording of entering and exiting vehicles. Initial results show significant improvements in maintaining traffic order and safety, as well as facilitating residents and visitors in finding parking spots.

Indoor/Outdoor Parking Lots: This system helps manage and monitor parking areas efficiently, reducing illegal parking and optimizing the use of parking spaces.

Smart City Development Projects: This technology is an important part of smart city projects, helping to improve traffic infrastructure, reduce congestion, and enhance the quality of life for residents.

Apartment Renovation Projects: In apartment renovation projects, the parking control system ensures that residents have enough parking spaces while minimizing the impact on the surrounding area.

Large Vehicle Parking Lots: This technology is also applied to parking lots for large vehicles, such as buses and trucks, helping to manage parking spaces and logistics process more efficiently and safely.

Parking Control Technology is being widely applied in various fields and projects

Awards and Recognition

icon award

LightVision Inc. has been honored as the best company in the IMP (Idea Market Place) program, a prestigious competition for domestic and international technology companies organized by POSCO. This is an important recognition, demonstrating the company's excellence and development potential in the technology field.

icon award

Additionally, the company has received direct investment from the POSCO IMP Fund and Credit Guarantee Fund proving POSCO's trust and commitment to supporting the sustainable development and innovation of the company.

icon specification

LV Intelligence Specifications

Camera Support 2 CCTV cameras per GPU (based on 15 fps speed)
Camera Resolution FHD (1080p)
Network Environment 1 Gbps
Video Storage Device NVR
Vehicle Detection Speed under 30 km/h (performance can be improved)
Types of Vehicles Detected small cars, large cars, forklifts, tankers, trucks (small, large), sports cars, etc.
Recognition Performance 30 milliseconds/frame
PC Requirements Intel Core-i5, 32GB RAM
GPU Requirements RTX-3060 GPU with 12GB RAM