Evaluating Machine Vision Cameras versus Comparing Camera Specification Documents

Posted by Gretchen Alper on Wed, Jan 30, 2013

Many of our customers ask us how they should evaluate cameras in order to find the best solution for their application. There are several reasons why we feel the Photon Transfer Curve (PTC) is useful, especially if you are trying to get the most out of your imaging system such as when detecting small differences between images or using the pixel information as data for measurements. We have shared some detailed information on how to measure the Photon Transfer Curve for CCD or CMOS machine vision cameras because we feel it is an important part of camera evaluation and selection.

Camera vendors provide a lot of detailed specifications about their cameras, but as we have discussed it is challenging to connect all of the camera specifications to the image details you require and to know how the camera will perform under your system settings.  The biggest advantage of measuring the Photon Transfer Curve of different cameras yourself is that you can verify the camera performance with your own equipment and environment such as integration time, illumination conditions (continuous/strobed, spectral range), optics, filters, etc. which can all have an effect on the performance.  You are then not only comparing cameras under the same conditions, but also comparing them under conditions that are relevant to your system/application.

Should you have some questions about how to perform the measurements, there are many resources or courses available. 

As you could read in an earlier blog, the EMVA 1288 standard provides details on how to accurately measure many camera parameters in a controlled environment.  This can be very useful for camera manufacturers as guidelines to present specifications or for qualifying cameras during development. (Note that the Photon Transfer Curve is measured by EMVA 1288 too!)

But instead of using the EMVA 1288 test setup, you can also use your own system to gather the measurement data needed to construct the PTC and derive important camera parameters like full well capacity, read noise, dynamic range and linearity.

If you are a customer that wants to get the most out of your system, we strongly suggest that you invest in testing and understanding camera performance.

Testing of a camera using your specific imaging conditions may show behavior and image artifacts that are not seen under "standard" conditions that are used during camera testing in the factory of the camera vendor. In practice we've often identified significant differences in camera behavior from different vendors - even between cameras that are using the same image sensor!

There is no camera in the world that that will be an optimal fit for your application.  Your competitive advantage depends on your electro-optical system, measurements algoirthm, etc. being unique.  Therefore, to summarize the top 3 reasons for measuring the PTC are so you can:

  • Verify and compare camera performance under the relevant system conditions.
  • Better communicate questions to the supplier or understand customization requirements.
  • Go beyond megapixels/second to quantify image quality to parameterize the outgoing inspection at the supplier or incoming inspection needed to have consistency in series production. 

For more details on comparing cameras, please see our free ePaper or our local office would be happy to support you.


Related Blogs:

5 Steps to Prepare for a Machine Vision Camera Evaluation 

A Few More Steps for Machine Vision Camera Evaluation and Selection

How to measure the Photon Transfer Curve for CCD or CMOS cameras

How to measure the Photon Transfer Curve for CCD or CMOS cameras (2)

Does the EMVA 1288 Standard help you decide on the right machine vision camera for your application?

Topics: Vision System Optimization

Previous blog:

Next blog: