HomeChipsetsSoftBank’s ARM says ADAS camera can outperform human eye

SoftBank’s ARM says ADAS camera can outperform human eye

The chip technology architects at ARM, which was acquired by Japan’s SoftBank last year for $32 billion, have created an image signal processor that targets advanced driver assistance systems. The company said ADAS will require not only more cameras inside vehicles, but more powerful cameras.

“Simply integrating standard camera technologies used in smartphones, or even consumer video cameras, is not an option,” said Tim Ramsdale, general manager for image and vision at ARM, in a blog post. “Complex camera technology must quickly process and analyze images under the most extreme conditions and be specifically designed for stringent automotive safety standards.”

ARM’s response is the Mali-C17 image signal processor. The chip is able to process up to 4 real-time cameras and 16 camera streams with a single pipeline.
It has more than 300 dedicated fault-detection circuits and is capable of handling 24 stops of dynamic range. It can also capture twice the dynamic range of a standard single exposure sensor. ARM said that in some cases the Mali-C17 can outperform the human eye.

“Automotive cameras need to provide the highest possible levels of clarity and reliability,” said Ramsdale. “Where the camera output is generated for a display, image quality is paramount. Take for example mirror replacement … The Mali-C71 offers ultra-wide dynamic range (UWDR) up to 24 stops (in some cases, providing detail beyond what the human eye can see).”

Vehicular image signal processors have two important jobs. They need to capture camera data and generate images for the driver to see on a display, and they need to simultaneously process that data for the car’s computer vision system. Ramsdale said these two tasks create conflicting requirements for the chip, but nonetheless ARM has created one piece of hardware intellectual property that can handle both tasks.

Previous post
Huawei unveils LTE-based smart grid solution
cloud RAN microsoft
Next post
Schneider Electric teams with Microsoft to accelerate the IoT