RAW compression and TICO-XS about to revolutionize industrial vision

11.12.18 02:35 PM By Nils Finger

Image sensors are at the heart of every vision system

Just like in every part of our daily life, video is also becoming omnipresent in industrial applications. With better machine learning and AI, the output from image sensors becomes more and more relevant. Higher speeds and greater resolutions help optimize industrial vision and autonomous driving systems. However, with more, faster and higher quality video, data amounts are rising significantly and become hard to handle.

intoPIX could be a game changer in this situation. Including TICO RAW right at the image sensor means completely rethinking the way we transport video from capture to its final usage. Unlike in current applications, the video signal is compressed in its RAW format before destructive image processing takes place. This reduces file sizes straight from the sensor, thus reducing overall power consumption and increasing the speed at which this image data can be transported and analysed.

TICO RAW compression specifically excels due to its low power processing, lossless quality properties and ultra-low latency, making it invaluable for good analytics. 

JPEG XS replacing JPEG and uncompressed video 

in VISION applications

The upcoming image coding standard JPEG XS has already been causing excitement across the broadcast and consumer electronics industry. In both industries it is seen as an alternative and improvement over systems that currently make use of uncompressed video. intoPIX now introduced the novel technology to the industrial vision market at this year’s VISION Show in Stuttgart.

Industrial vision technologists also see great potential in future TICO-XS IP cores (standardized at ISO as JPEG XS) as a replacement for the currently widely adopted JPEG standard. Due to JPEG XS’ ultra-low latency of only a few microseconds and remarkably smaller logic for implementation, but also the vastly increased color bit depth (16 bit instead of 8 bit), the already 25 year old “conventional” JPEG technology has found a potential successor.