Smart cars are not only a typical scenario of digital-real integration, but also a vanguard of digital transformation. Among them, the diverse, massive and rich AI data is the foundation for the stable development of autonomous driving. Together with other key technologies, it is driving the arrival of the era of full commercialization of autonomous driving step by step.
On November 4th, at the “2021 Tencent Digital Ecology Conference Smart Travel Special”, Tencent released the “Autonomous Driving Cloud Platform”. With its rich and mature AI data service capabilities and experience in the field of smart cars/autonomous driving, Cloud Test Data has become an ecological partner of Tencent’s autonomous driving cloud, and has cooperated with Siemens Industrial Software, Hexagon, IPG China, China Automotive Engineering Research Institute, Shanghai Zhuoyu, Cognata, Zhejiang Tianxingjian and many other industry representatives will jointly promote the development of autonomous driving technology and create new value in the travel industry chain.
At present, the digital transformation of the automobile industry has entered the “deep water area”. On the one hand, car companies must reconstruct their business systems and establish digital channels for all links of the entire industry chain. On the other hand, they must also build a user-driven service operation model to explore new business value. Behind this, the integration of digital technology is required to create new value in the travel industry chain.
In the smart travel scenario, the training data requirements of the entire process from the pre-research period of the relevant algorithm to the project implementation are different. The required data includes many sub-scenarios such as DMS and ADAS. Come higher, including road object labeling, intent recognition, 3D point cloud labeling, 2D/3D fusion labeling, panoramic semantic segmentation and many other labeling types, which are very important for the software engineering capabilities, system capabilities, and collaborative labeling capabilities of different dimensional data of the labeling platform. presented a challenge.
Cloud Testing Data and Tencent Autonomous Driving Cloud have launched an ecological cooperation. The two parties will work together to deeply integrate the training data labeling technology in the field of intelligent driving to realize the connection of data return, preprocessing, labeling, and output, so as to meet the needs of rapid iteration of algorithm models and help intelligent driving. The high-quality production of training data in the field of driving effectively connects all aspects of the entire life cycle of autonomous driving research and development, including data collection, storage, labeling, algorithm training, simulation, evaluation, mass production data return, and data operation. . As one of Tencent’s first ecological partners, Yuntest Data, together with other partners, focuses on providing full-link services for autonomous driving technology research and development to support the industry in more efficient autonomous driving research and development and operations.
The cloud measurement data labeling platform is a full-process processing tool for AI data required in the actual production of enterprise AI training data. Through structured, intelligent, engineered and standardized research and development, it covers data collection, data labeling, data management and other data processing links.
At the technical level, the cloud testing data labeling platform has the advantages of multi-terminal data support, AI-assisted quality inspection, rich labeling tool support, streamlined and efficient operation, deep integration of enterprise processes, and quality control of labeling processes. Version management, annotation results visualization and other functions can meet the data requirements of the diversity and richness of AI landing scenarios, and the overall efficiency of the AI data training process has been improved by 200%; at the tool level, the cloud measurement data annotation platform supports images, text, voice, and video. One-stop processing and processing of data types such as point clouds, with 3D solid frame, point cloud semantic segmentation, feature points, line segments, rectangular frames, curves, plane solid frames, polygons and other types of professional tool components required in the industry, flexible Meet different labeling needs, cooperate with algorithm models for data processing, and quickly respond to the diverse needs of AI training.
In the future, with the continuous deepening of cooperation between the two parties, on the one hand, it will bring more professional and richer data annotation platform tools to smart driving users, and on the other hand, it will attract industry-leading autonomous driving development tools and products to jointly build an industry ecosystem and provide smart mobility. , AI companies provide more convenient and smarter digital platforms to accelerate the development of autonomous driving technology.
The Links: 7MBP75RA060 FF300R17KE3