Is Higher Lens Magnification Always Better?

 

I recently received this question: "Why does your vision system have only a 2.7x magnification when these other microscopes offer 20-40x or higher magnification?" 

Lens magnification must be balanced with the operation and performance of the system for the desired task.  Bigger is not always better.  In addition, lens resolvability must be taken into account as well.  If a lens can magnify an object 100x but only provide resolvability 1/10 of this, the effect is 10x of un-resolvable features or essentially a blurry image.  This is a common mistake in vision system applications and why specifying a lens appropriately is so important.  

The magnification of a lens can be calculated by the size of the object and the size of the camera sensor  (although, in actuality, it is independent of sensor size) as Mag = imager_size / object_size. Therefore, if the object we want to measure is 1.3mm (and say we have a reasonable field-of-view (FOV) that is 20% larger than this, so 1.3 *1.2 = ~1.5mm) and our sensor height is 10.4mm (vert dim of a 1.1” diag sensor for 12MP camera), then the max magnification we would want to use in this instance is 10.4 / 1.5 = 6.6x. A FOV of 1.5mm might allow us to measure only certain features on the object of interest. If we wanted to measure smaller features on that object at a higher magnification, we could do so but our FOV would additionally be much smaller (we might only see one small portion of the object!).

Lens magnification must be balanced with the task at hand.  Is a larger FOV required?  If so, it will be at the expense of magnification.  In the situation that produced the question that prompted this post, we chose to simultaneously measure several other features of the object in a single image, which required a larger FOV.  Although we enlarged the FOV, our goal was to maintain measurement repeatability within the required tolerances.  If we had instead chosen to go with a higher magnification, obtaining the other measurements would have required adding servos/micrometer stages to move the camera and/or object so that the camera can see different parts of the object at the higher mag (smaller FOV).  While that was certainly an option, it also would have taken significantly longer processing time for our application.  It is also for reasons like this that gauging applications should have reasonable and realistic tolerance requirements specified. I have worked on many applications where arbitrarily very tight tolerances were initially required and would have required significantly more complex and expensive gauges to meet such specifications.  In these situations, it is important to agree upon reasonable requirements which can potentially save a lot of time and money. 

Lens magnification is simply another variable that must be balanced with several others to achieve successful results for a given task.