Ambarella, a developer of video processing and computer vision semiconductors, and OFilm, a provider of intelligent driving solutions, announced a collaboration on an automated valet parking (AVP) hardware platform.
The platform uses Ambarella's CV22AQ CVflow computer vision system-on-chip (SoC), which, according to the company, features superior image signal processing, powerful artificial intelligence (AI) computing performance, and ultra-low power consumption (typically under 2.5 W). That is used in conjunction with OFilm's intelligent driving algorithms and system integration capability. The companies say that the platform achieves high-precision image recognition, free space detection, obstacle detection, and map construction.
"OFilm continues to address customers' needs and is committed to building intelligent driving technology and products that can be mass produced," said Dr. Ma Guanglin, General Manager of Intelligent Driving at OFilm. "Such products need powerful algorithmic capabilities combined with the ability to optimize for high performance. We chose Ambarella's CV22AQ SoC, which is based on their CVflow architecture, as it provided these capabilities."
"We are pleased to be working with OFilm Intelligent Driving to provide advanced automated valet parking platforms," said Ambarella CEO Fermi Wang. "Our CV22AQ CVflow computer vision SoC running OFilm's advanced AI software will enable the next generation of intelligent self-parking vehicles."
The CV22AQ is manufactured in an advanced 10-nanometer process, providing the ultra-low power consumption required for the design of compact automotive systems. Its CVflow architecture reportedly delivers real-time processing of up to 8 MP resolution video at 30 frames per second (fps) for high-precision deep-learning-based object recognition. The CV22AQ also supports multiple camera inputs with simultaneous output in multiple resolutions and viewing angles.
The CV22AQ's high-performance image signal processor (ISP) is designe to delivers superior image quality in low-light environments, while high dynamic range (HDR) processing extracts more image detail in high-contrast scenes to further enhance the computer's computer vision capabilities. It delivers network security features such as secure boot, TrustZone, and I/O virtualization while also supporting secure over-the-air (OTA) software upgrades.