Thursday, April 21, 2016

Automotive Image Quality Standard Proposed

Robert Stead, the founder of Sense Media organizing AutoSens conference, teams up with Patrick Denny of Valeo, Sven Fleck at SmartSurv, Benjamin May at AMX13, and a number of other professionals to submit IEEE Project Authorisation Request (PAR), Proposed Project 2020: Automotive System Image Quality.

From the submitted PAR:

5.2 Scope: This standard addresses the fundamental attributes that contribute to image and quality for automotive Advanced Driver Assistance Systems (ADAS) applications, as well as identifying existing metrics and other useful information relating to these attributes. It defines a standardized suite of objective and subjective test methods for measuring automotive camera image quality attributes, and it specifies tools and test methods to facilitate standards-based communication and comparison among OEM and Tier 1 system integrators and component vendors regarding automotive ADAS image quality.

5.4 Purpose: This standard specifies methods and metrics for measuring and testing automotive image quality to ensure consistency and create cross-industry reference points.


  1. Maybe they target the complete optical path vs. the sensor/electrical part that is covered with EMVA1288?

    1. This is already covered by ISO. Both standards can be combined to have a complete test.

    2. To my knowledge ISO & EMVA1288 together do not (yet) cover all aspects of image sensors/cameras. I know that you guys are working on a new EMVA1288 to also cover e.g. global shutter metrics, but there are still more gaps left, no? To my (possibly limited) knowledge/overview neither EMVA1288 nor ISO cover blooming or charge transfer (not lag - important for lock-in imaging) characterization. Also, many metrics are speed and/or irradiance dependent (noise/PRNU/DSNU/FWC) - but the EMVA1288 gives the user full flexibility in using either irradiance or integration time variation. This assumes that the product of irradiance and integration time is constant, which sometimes is a major simplification. The ISO standards are sometimes more specific, but nobody seems to be using those standards accurately (looking e.g. in direction of ISO sensitivity). Also there is this entire world around QIS, SPADs and SiPMs.

      So I think there's lots to do and I appreciate any means of standardization. If that can be done in an extended version of the EMVA1288, a new ISO standard or an entirely new standard by yet another community can, of course, be argued about. What's is actually the motivation of that standard? Being more specific than e.g. ISO/EMVA1288? Filling the missing optoelectronic gaps? Or is it maybe about other parameters that are relevant for automotive such as reliability and robustness?

  2. The EMVA1288 is for a sensor only. The ISO 14524/15739/15781/17321 and so on is for a complete camera system, including lens.


All comments are moderated to avoid spam.