Ambarella's CV5 Chip Boasts 8K Image Processing, AI Acceleration, and Encodes in Just 2W
Targeting AI cameras and supporting one 8K or four 4K video streams, the new part can encode 8K video in just 2W of power.
Computer vision specialist Ambarella has announced the launch of the CV5, a high-performance vision processor for artificial intelligence applications capable of processing a single 8K or four 4K video streams — and in under two watts of power.
"With the introduction of CV5, Ambarella is defining the next generation of automotive, consumer, and robotic cameras,” claims Fermi Wang, chief executive of Ambarella, of the company's latest launch. "By combining 8K single-channel and 4K multi-channel recording with the high performance of our CVflow AI engine, we are enabling cameras with the highest-quality imaging and innovative new AI features."
The CV5 system-on-chip (SoC) combines two Arm Cortex-A76 general-purpose processing cores with Ambarella's in-house CVflow AI vision acceleration engine and an image signal processor (ISP) which processes images for human and machine consumption simultaneously. According to the company's internal testing, the chip — fabricated on a cutting-edge 5nm process node — draws under 2W of power during an 8k30 video encode.
The company is targeting a range of markets with the new part, including automotive with the promise of support for multiple video streams including ADAS, driver monitoring, cabin monitoring, and side-view inputs, virtual reality camera systems, robotics and drones with support for accelerating simultaneous localisation and mapping (SMAL), path planning, obstacle detection and avoidance operations.
The chip is supported by the company's existing software development ecosystem, using the same toolchain as its earlier releases including the low-cost edge-AI-focused CV28M released late last year. The company has not, however, released pricing information.
More details on the part can be found on the Ambarella website.