Modulation Transfer Function

MTF stands for Modulation Transfer Function and corresponds to the measurement of the perceivable contrast and sharpness by an image detector. The obtained value is a function that depends on the frequency, wavelength and intensity of the observed features in the image. The MTF of an ideal detector would be constant, while real-life detectors show a decreasing trend in contrast as frequency increases.

MTF is a measure that is mainly used in photography, and most existing software focus on this use.
In the context of optical glass parts inspection, the MTF is rather used to characterize the effect of the glass on the image contrast and sharpness. Snellium offers a solution tailored for the latter, called sharpness or MTF reduction.

Slanted edge illustration

Slanted Edge Method

In our solution, the MTF is calculated using the slanted edge method, a commonly used method to approximate the MTF curve along the frequency spectrum of the tested camera system.
See the related ISO standard.

The scientific community did not converge yet to define intermediary operations as image sampling or smoothing methods. As consequence, depending on the experimental conditions, the user may need to adjust these values which, unlike other commercial solutions, is possible with our software.

MTF example for distorted image

Data Mapping

Our algorithm implementation is able to retrieve 2D maps of MTF results, giving a representation of the MTF variation in the whole field of view of the camera. Depending on the type of analysis and preference of the user, the various plots can be expressed at fixed contrast, frequency or as the integral of the MTF curve.

Moreover, the signals are split in horizontal/vertical or sagittal/meridional directions, allowing the user to choose the best view depending on the type of pattern and analysis. The software also gives the possibility to the user to configure the more appropriate mapping interpolation method.

Relative MTF example

Relative Measurements

The MTF measurement is affected by the glass, but also by the optical system itself: camera, lens, etc. To isolate the impact of the windshield, we do a comparison of two MTF maps: one from an image acquired through the windshield and one without it. The obtained data is called the MTF reduction.

From the comparison of the two measurements, we can deduce the relative effect that the windshield induced between the two images acquisition.

Unlike other commercial tools, our software features implemented in the slanted edge algorithm takes this objective in mind, and apply corrections smartly to avoid altering this relative measurement.

MTF pattern examples

Pattern Support

A lot of different patterns exist out there to do MTF measurements. Our algorithm detects edges smartly, making it virtually suitable for any kind of pattern you might want to use.

In addition, we propose our own pattern, specifically tailored for the ADAS region. The use of it provides multiple measures with a single image, increasing the productivity of your test campaign. The applicable algorithms include Digital Image Correlation and barrel distortion calibration.