Matching lenses with high performance cameras (high speed combined with high resolution)

Posted by Gretchen Alper on Thu, Oct 13, 2011

To get the full performance of high speed and high resolution cameras, the appropriate high-performance optics are required.

High resolution used to mean high priced industrial cameras.  With the improved performance of CMOS sensors, high resolution cameras are more cost-effective and attainable for more machine vision systems.  A challenge for system builders is that these higher resolution cameras also have a smaller pixel pitch.  While the sensor size and therefore optical format may be the same as the existing camera in the system, simply upgrading to a higher resolution camera may not provide increased image details. Using the existing lens and mechanical set-up may produce an image with less sharpness and contrast due to the higher spatial resolution of the image sensor/camera compared to that of the lens.

As this recent article from Vision Systems Design nicely describes, increased resolution cameras with the same depth of field of the system will require a different lens to fully realize the advantages of higher resolution cameras.  Lower aperture lenses or more complex lenses are often necessary.

Vision Systems Design Article: Depth of field is crucial to maximizing performance in camera systems by Gregory Hollows

We experienced this challenge ourselves with our camera lens assembly that is designed for digital IITV X-ray chains.  While we maintained the resolution, the change to smaller pixel pitch and higher spatial resolution affected the overall system design.  “When designing our Pearl cameras we switched the sensor from the KAI-1020 with 7.4 um square pixels to the KAI-01050Adimec CLA Pearl high resolution with 5.5 um square pixels. As a consequence, the maximum spatial frequency at the sensor for a similar spatial frequency in the object space (the image on the II phosphor screen) increased from 68 to 91 lp/mm.  This resulted in an increase in lens complexity for identical performance. A change in complexity means e.g. an additional lens element as in case of Pearl, but could also mean the application of aspherical components. The latter is still rather expensive for small/medium series systems.” René Aartsen, Electro-Optical System Designer at Adimec.


The latest higher resolution CMOS cameras also have higher frame speeds which mean lower integration times, again resulting in extra demands on the lens.  More complex optics could be needed to get enough light to the sensor, but this depends on the high resolution camera that is selected.  Using a camera with high resolution, high speed, and full optimization of the sensor can simplify the required lens related to sensitivity.  A camera with lower read noise (e.g. better sensitivity) is more capable of detecting small intensity variations at lower light levels.  The overall sensitivity of the camera and the lens needs to be considered.

Enhanced higher resolution cameras can offer additional benefits to the overall system, such as the ability to store lens calibrations in the camera and advanced sensor corrections and flat field corrections.  The "simple" correction algorithms that are used for lower resolution cameras will no longer be sufficient at the edge of the image.  

The conclusion is that one has to re-evaluate the complete electro-optical system in order to maximize the ROI. More resolution and speed leads to higher quality lenses which leads to higher quality cameras with the latest image enhancement and processing. Otherwise the full benefit of the investment in upgrading the system is not realized.

Topics: Vision System Optimization, Optical Enhancements