ISO 16505 — Road Vehicles — Camera Monitor Systems

 

News

We will be presenting a paper, “Measuring MTF with wedges: pitfalls and best practices” at the Multi-camera and Embedded Systems for Autonomous Machines session of the Electronic Imaging Conference at San Francisco Airport on January 30, 2016. (It was a late submission, so it may not be in the program yet.) The paper focuses on difficulties with the ISO 16505 standard and how to overcome them— how to obtain good measurements despite ambiguities in the standard, which is in urgent need of revision.

A key finding in the paper is that the primary resolution metric in ISO 16505, MTF10 derived from hyperbolic wedges, is not a reliable measurement of system performance. A poor quality camera can have a good MTF10 measurement. We recommend replacing MTF10 with the minimum of {MTF10, the Nyquist frequency, and the onset of aliasing (the spatial frequency where the bar count starts decreasing)}.

We will be adding ISO 16505 measurements tips on this page.

Test Targets

ISO-16505 Test Targets are available for purchase on the Imatest Store

Test Methods

Resolution

references ISO 12233:2014.   Hyperbolic wedge features can be analyzed by the Wedge or eSFR ISO module. Slanted edge features can be analyzed by the SFR, SFRplus, eSFR ISO, SFRreg , or Chekerboard modules.

According to the standard, the contrast of the hyperbolic wedge should be greater than 20:1 or 40:1 (depending on where you look) so some lower contrast wedge test charts will not strictly comply with the standard. Very high contrast charts have problems with saturation or clipping.

The standard has some confusing and possibly erroneous statements on spatial frequency units.

  • Imatest calculates spatial frequencies directly from the wedge geometry itself. It does not use the numeric markings next to the wedges in any of its calculations. These numbers can be multiplied by 100 to obtain spatial frequency in Line Widths per Picture Height (LW/PH) if the image is framed so the top and bottom borders of the chart, which contain arrow marks (∇ and Δ), are at the exact edges of the frame. The problem is that charts are rarely framed this way in practice.
  • The fundamental units used by Imatest to calculate spatial frequency response are cycles per pixel (C/P), where 1 cycle = 1 line pair = 2 line widths.
  • Spatial frequency in Line Widths per Picture Height (LW/PH) is normally obtained by multiplying the C/P value by 2 * the Picture Height in pixels (typically the shortest dimension of the total picture).
  • ISO 16505 specifies a hypothetical square region inside the image for calculating LW/PH spatial frequencies. If Lpx is the length in pixels of a side of the square, then you can use the simple equation, LW/PH = 2 * Lpx * C/P. The standard has many confusing statements about units. This is most of what you need to know.

The units LW/PH can be selected in the settings windows, and a secondary readout of MTF10 can be selected to produce a compliant measurement.

Magnification

The ISO 16505 requires magnification to be mapped as a function of field distance. Imatest currently computes optical distortion as a function of field distance, but does not yet convert this into a magnification output.

 Aspect Ratio

The standard refers to aspect ratio across the field, and where that aspect ratio is at a minimum value. Imatest doest not currently consider aspect ratio as a function of field distance and instead computes an overall aspect ratio.

Color

The patch pattern of the ISO-16505 is available in multicharts & multitest.

The CIE-1976 degree-based color output is also on our short term roadmap.

Timing

There is currently no calculation of system latency produced by Imatest.

 See Also