Measurement of Machine Vision Image Sensor/Camera Parameters

Posted by Gretchen Alper on Thu, Nov 3, 2011

Characterizing the image sensor/deciding on the image sensor is an important first step in camera selection for a machine vision application, but it is not a simple task.  There are some courses such as the one highlighted below that can offer excellent training and support.  You can also contact Adimec for information as we have thoroughly evaluated image sensors we use in our cameras as well as many we don’t.  Check out our white paper on comparing different cameras when you have selected a desired image sensor.

Instructor Albert Theuwissen,-albert-jp-.aspx


Course Name:  Hands-on Characterization of Solid-State Image Sensors,-speech-and-signal-processing/c04-hands-on-characterization-of-solid-state-image-sensors.aspx

Course Description:

Measurements in Dark
It is astonishing how many parameters of a sensor/camera that can be measured without any light input. The photodiodes of the image sensors are collecting charges and in principle it does not matter whether these charges are being generated by photons or whether they come through leakage in the structure. Once a charge packet is collected in a photodiode, it can be used to characterize the imager!
By means of dedicated dark measurements the following parameters can be evaluated:

  • Dark Current Density
  • Dark-current Non-uniformities
  • Linearity
  • Conversion Gain
  • Dark Reference Lines/Pixels
  • Leaking Pixels
  • Temporal Noise
  • Dynamic Range

Measurements with Light
Image sensors and cameras are being developed to convert incoming photons into a measurable quantity. Therefore, it is also of great importance to evaluate the light characteristics of the devices. Parameters that can be measured when photons are impinging on the sensor are:

  • Light sensitivity
  • Light non-uniformities
  • Linearity
  • Conversion gain
  • Blooming characteristics
  • Black sun
  • Green-green differences
  • Reciprocity effect
  • Saturation level

Different participant subgroups will work on different set-ups, not necessarily performing the same measurements. However, at the end of the course, everyone will get copies of the outcome of all characterizations performed by the groups.

describe the image

Topics: Vision System Optimization

Previous blog:

Next blog: