LiangDao Intelligence partners with Japanese Tier-1 supplier for BEV data training

Beijing (Gasgoo)- Recently, smart LiDAR solution developer LiangDao Intelligence announced a data business partnership with a globally renowned Japanese Tier-1 supplier.

Specifically, LiangDao Intelligence will provide high-quality 4D time series true value data combining BEV vision and LiDAR point cloud to support advanced BEV perception algorithm training and functionality development for customers. This marks LiangDao Intelligence's first large-scale delivery of 4D time series true value data to a Japanese client.

Photo credit: LiangDao Intelligence

Developing intelligent driving features requires continuous optimization of perception algorithms, which depend on the efficacy of data feedback loops and the volume of training data. Additionally, the challenge of storing, managing, and seamlessly integrating vast amounts of data across multi-cloud platforms and localized R&D systems is a significant hurdle for many global automakers.

Leveraging its extensive experience in intelligent driving perception algorithm development and application, LiangDao Intelligence has introduced a comprehensive data solution for in-vehicle perception training. This includes global one-stop data services, high-quality intelligent data annotation, a cloud-based big data management platform, and proprietary intelligent data processing toolchains. These services and products are highly automated, significantly boosting production efficiency while ensuring data compliance and security.

At the core of the company's perception training toolchain is the continuously optimized perception algorithm, tailored to meet individual customer needs. LiangDao Intelligence's collaboration with the Japanese Tier-1 supplier is attributed to the recently launched BEV vision + LiDAR point cloud fusion data offline algorithm. This algorithm's efficient data processing and high automation levels enabled LiangDao to stand out in the global supplier evaluation for this project.