Facing the $2.4 billion  forward-looking camera market, major enterprises compete to carry out new projects, focus on delivering comprehensive solutions and help automobile manufacturers speed up the process of product listing
On September 1, 2021, Shanghai, China – Xilinx, Inc. (NASDAQ: xlnx), a leading enterprise in adaptive computing, and motovis, an embedded AI autopilot provider, today announced that they are cooperating to launch a solution for the automotive market, which will be Xilinx car specification (XA) zynq ® The system on chip (SOC) platform is combined with magic vision intelligent convolutional neural network (CNN) IP, which is specially used for vehicle perception and control of front view camera system. The solution relies on Xilinx’s company plan and aims to provide customers with a strong platform to strengthen and accelerate development.
Figure: Xilinx and magic vision jointly launched solutions for the automotive market to help the innovation of front-view cameras
The front view camera system is a key element of advanced driving assistance system (ADAS) because it can provide the required advanced perception for safety critical functions, including lane keeping assistance (LKA), automatic emergency braking (AEB) and adaptive cruise control (ACC). This now on the market solution uses convolutional neural network to realize a cost-effective combination of low delay image processing, flexibility and scalability, so as to support a series of parameters required by the 2022 European new car assessment program (NCAP).
Ian riches, vice president of global automotive business (GAP) of strategy analytics, said: “This cooperation is an important milestone in the forward-looking camera market, because it will enable automotive OEMs to innovate faster. The forward-looking camera market has great growth opportunities. We expect annual shipments to increase by nearly 20% year-on-year from 2020 to 2025. Xilinx and magic vision will work together to bring a highly optimized software and hardware solution, which can greatly improve the market Meet the needs of automotive OEM manufacturers, especially with the emergence of new standards and growing demand. “
The forward-looking camera solution is extended by magic vision’s CNN IP across 28nm and 16nm vehicle gauge zynq SOC series devices. This unique combination of optimized software and hardware function division and customizable CNN special engine based on magic vision intelligent deep learning network provides cost-effective solutions at different performance levels and price points. The solution supports image resolutions of up to 8 megapixels. Now, OEM manufacturers and Tier-1 suppliers have superimposed their own unique algorithms on the perception algorithm set of magic vision intelligence for the first time, so as to bring differentiated and future oriented design.
Dr. Yu Zhenghua, CEO of magic vision intelligence, said: “We are very pleased to launch this new solution together with Xilinx and bring our CNN forward-looking camera solution to the market. When designing a system with AEB and LKA functions, customers need to carry out efficient neural network processing in SoC so that they can flexibly and easily realize future functions. Magic vision has a customizable deep learning network, and The Xilinx zynq platform has the ability to support CNN’s dedicated engine. With the unparalleled efficiency and optimization provided by the two, we are helping to build future oriented design with the goal of meeting customer needs. “
Market forces continue to promote the adoption of forward-looking camera system to comply with global government mandatory requirements and the expectations of consumer observation groups, including the general safety regulations of the European Commission, the national highway traffic safety administration and the European new car assessment program. All three parties have issued formal injunctions or strict guidelines to require automobile manufacturers to realize LKA and AEB functions in new vehicles produced in 2020-2025 and beyond.
Willard Tu, senior director of Xilinx automotive business, said: “With a comprehensive solution for the forward-looking camera market, we have expanded the lineup of vehicle specification level solutions and delivered cost optimized high-performance solutions to customers. We are happy to put this solution on the market and promote the development of the industry. Magic vision intelligence has accumulated rich expertise in the field of embedded in-depth learning and can be optimized Through the network to cope with the severe challenges perceived by the front-view camera, these provide unique advantages for both sides to win market share, and will also help our OEM customers speed up the time to market. “
Xilinx and magic vision will speak together at Xilinx adapt 2021 virtual event held on September 15, 2021 us time. Xilinx adapt 2021 includes high-level keynote speeches and practical statements from partners and customers to help users unlock the value of adaptive computing. Xilinx adapt China station series activities will be held online in November 2021. For details, please pay attention to Xilinx Chinese website or Xilinx official wechat.