July 26, 2022

XfuseAI Board400x275Xfuse LLC, which develops customizable imaging and video technology, has announced it intends to deliver hardware and software tools to advance the next-generation of autonomous vehicles, robotics, and artificial intelligence applications. The company’s proprietary Phoenix HDR Image Signal Processing (ISP) technology supports multiple different sensor types simultaneously to fuse data-rich HDR video in real time, with minimal latency.

The company said the location-aware data streaming technologies offered by the Xfuse Phoenix HDR ISP offer tools to help rapidly advance self-aware robotics, autonomous vehicles, and machine vision applications. Video data that can be fused through the system include  output from multiple non-image sensors, such as thermal, GPS, lidar, radar, gyroscope, as well as other Xfuse software and tools.

“Our proprietary ISP offers engineers full control over the complete imaging pipeline to fuse data streaming from an array of imaging and non-imaging sensors in real time,” said Alfred Zee, president and CEO of Xfuse. “The market for autonomous mobility of all types is currently dominated by a few very large companies and is expected to grow to $675 billion by 2030. By offering our technology to both in-house development teams and independent developers, we are helping them streamline their time to market and lowering development costs for the next generation of autonomous vision applications.”

The Phoenix HDR ISP is a fully programmable ISP that can be customized to each developer’s needs. It can process imaging and non-imaging data from numerous sensors regardless of brand, resolution, or color filter array. The Phoenix ISP can process video with extremely low latency and new external DRAM or frame buffers, Xfuse said. It runs exclusively in the FPGA fabric to free on-chip CPU and GPU computing for improved AI functionality.

The company said it will roll out the Phoenix HDR ISP to developer platforms soon. For more details on the product, visit the Xfuse website here.