sevice@vcarsystem

*Name

*Mobile

+86

*E-Mail

*Company

*Support types

*Country

*Province/City

*District

Leave Message

I agree with the PrivacyPolicy and terms and conditions of VCARSYSTEM
In autonomous driving R&D and testing, data acquisition and reinjection have been widely applied. As a key device in road data collection, the YUV camera captures image data that is crucial for technological iteration. However, YUV format data has a large volume and high storage and transmission costs. Current processing methods are limited, especially in terms of efficiency when extracting and applying ground truth information, which restricts the full potential of the data's value.

VCARSYSTEM's Solution

The YUV camera compression processing acquisition solution utilizes efficient compression techniques (preserving ground truth information and core details) to generate high-quality data that can be directly mined. This approach not only extracts ground truth labels for target objects to support model training, but also extracts ground truth from scenes such as intersections and ramps to build databases for development and testing. Additionally, it can be converted into virtual scenes containing ground truth to establish simulation libraries. Ultimately, this solution enables the efficient application of data across multiple scenarios, providing strong support for autonomous driving R&D and testing.

ADS 3000

The Industrial-grade AI Platform supports GMSL cameras, sensor data acquisition, and ADAS function evaluation.

WORKFLOW

FUNCTIONS

Adaptation and Support for Model Training

The solution is compatible with VCARSYSTEM's high-quality data acquisition, providing high-quality model training data to assist in precise model training.

Automated Label Generation

Enables automated label generation, effectively improving labeling efficiency and reducing manual labeling costs.

Object Perception Capability

The object perception is rich and accurate, providing a reliable basis for subsequent processing and decision-making.

High-Quality Data Acquisition

Multi-sensor synchronization accuracy ≤1ms, ensuring the quality of data acquisition and supporting sensor data fusion and utilization.

High Synchronization Exposure Control

Using self-developed FPGA for synchronized triggering of exposures across cameras, ensuring simultaneous data collection across cameras and maintaining data consistency.

Efficient Resource Utilization

Video acquisition is performed using a video capture card, with H.264/H.265 encoding and compression carried out at the board level, reducing hardware resource requirements and enhancing storage efficiency.

ISP Processing Support

ISP processing of camera RAW data at the board level, ensuring color restoration of video data and optimizing the overall video quality.

ASSOCIATED PRODUCTS

ADS 3000

More

Download Center

SOFTWARE

Download

VIDEOS

Download

USE CASES

Download

GET SUPPORT

Contact us

To address the growing complexity of R&D requirements, we will tailor solutions to your specifications and offer expert consulting and engineering services.

Contact us