EMVA 1288 Course, New Release 4.0 - Interactive Online Training

Time：Three-day course, December 01-03, 2020（13:00-18:00）

The 1288 standard of the European Machine Vision Association (EMVA) for objective camera characterization is adopted worldwide and makes it easier to compare cameras and image sensors of different vendors. In addition, the standard also evolved to an indispensable tool to shorten development cycles and improve the quality of cameras and sensors. The standard characterizes the devices strictly by physical parameters. It introduces objective criteria for the rating of sensitivity, noise, spectral sensitivity, dark current, nonuniformity and defective pixels. A one-page standardized summary data sheet makes comparisons easy. End of 2020 a new and significantly extended Release 4 will come into effect. The course will incorporate all major new features of this new release, including extended analysis of nonuniformity, non-linear (HDR) cameras, cameras with preprocessing, irradiation also through a lens, and various multimodal sensors (UV, SWIR, multispectral, polarization, time-of-flight).

Instructor

Prof. Dr. Bernd Jähne
HCI, Heidelberg University (https://hci.iwr.uni-heidelberg.de/dip)

Optional examination to be certified as an EMVA 1288 expert

Examination fee of 200 € for EMVA members and 250 € for non-members will be invoiced by EMVA.  Online test in English language lasts 90 minutes. Choose most suitable time in a period of about one week after the course. The required level of knowledge is taught in this course. A successfully passed expert level exam certifies that the holder has acquired all knowledge to perform EMVA 1288 measurements successfully and that he can interpret the measurment results correctly.

EMVA 1288 Course - Content Description of 3-Day Course：

Day 1 (three 90 minute sections, with two 15 minute breaks)

I. Introduction

• Purpose, history of EMVA 1288 standard

• Comparison with other standards, especially the ISO standards and the upcoming IEEE standards P2020 for automotive sensors and P4001 for hyperspectral imaging

II. Foundation Imaging and Image Sensors

• Radiometry of imaging; radiometric quantities

• Functional units of image sensors and digital cameras

• Inner photo effect; photon conversion in image sensors; noise sources

• Elementary processes in solid state imagers: optical losses, recombination, diffusion, and charge storage

• Causes and types of nonuniformity

• Dark current and its temperature dependency; temperature dependency of other sensor parameters

III. In-depth Foundation of the EMVA 1288 Standard

• “Black box” model for linear and non-linear cameras; system theoretical approach

• Characteristic curve and photon transfer curve

• Signal-to-noise ratio (SNR) for input and output signal

• Causes and types of spatial nonuniformity; separation into column, row and pixel variation; total signal-to-noise ratio

• Dark current and its temperature dependency; temperature dependency of other sensor parameters

Day 2 (three 90 minute sections, with two 15 minute breaks)

IV. Optimal EMVA 1288 Measurements and Measuring Conditions

• Mandatory and optional measurements

• Choice of analysis between linear model (Release 4 Linear) or general model (Release 4 General)

• Dark current measurements

• Irradiation series for sensitivity, noise, and linearity; light sources, irradiation variation, and irradiation without and with lens

• Measurement and analysis of nonuniformity

• Spectral measurements

V. Practical Exercises

• Performing EMVA 1288 measurements with available test equipment, software, and cameras

VI. Standardized EMVA 1288 Data Sheet

• Cover sheet

• One-page summary

• Main part with all EMVA 1288 measurements and graphs

Day 3 (three 90 minute sections, with two 15 minute breaks)

VII. Analysis of all Measuring Curves, Deviation from Ideal Behavior, Accuracy of Parameters

• Characteristic curve (sensitivity measurement); estimation of slope by cubic B-spline regression

• Photon transfer curve: a sensitive tool to analyze deficiencies

• Input SNR curve

• Problem of too low dark noise (influence of quantization noise)

• Derived parameters and their precision, accuracy and reliability: quantum efficiency, absolute sensitivity threshold, saturation capacity, maximum SNR, dynamic range (DR)

• Non-linearity

• Horizontal and vertical profiles to characterize nonuniformities and defect pixels in one representation

• Defect pixels: flexible characterization of defective pixels using logarithmic histograms

• Spectrogram method (Fourier transform to detect periodic nonuniformities)

• Dark current and its dependence on temperature

• Spectral sensitivity and color qualitative: measuring and evaluating the spectral sensitivity, characterization of color quality

VIII. Tests for Correctness of EMVA 1288 Measurements and Model

• Cumulative histogram

• Stability analysis

• Pixel correlation: is there any preprocessing (noise suppression etc.)?

• Dependency of PRNU on saturation

• Dependency of PRNU on wavelength of irradiation

IX. Application-oriented camera selection

• What do we need to know about the application?

• Which of the EMVA 1288 parameters are the most important given certain application conditions?

X. Outlook Further Development of the EMVA 1288 Standard (Release 4.1)

• Dependency of sensor sensitivity and nonuniformity on f-number;

• Spatial resolution of image sensors (PSF and MTF)

• Color quality