Magnification in microscopic imaging
At its core, microscopic imaging relies on magnification to produce an enlarged view of the sample. There are two types: optical magnification and digital magnification. However, because digital magnification (which uses interpolation) enlarges images without adding new detail and is unsuitable for precision inspection, this article will focus on optical magnification.
Optical magnification
Optical magnification physically enlarges the image before it reaches the sensor. This happens through the combined effect of the objective lens and the tube lens. When calculating system magnification, you can take one of the following two approaches depending on your setup—either standard manufacturer objectives or custom lens assemblies.
A. Using manufacturer specifications (standard systems)
In most industrial and research microscopes that use infinity-corrected optics, objective lenses are labeled with a nominal magnification (e.g., 10×, 20×, 50×). This magnification assumes the use of a standard tube lens focal length — typically 200 mm, though some manufacturers use 180 mm or 165 mm.
When building a vision system, use manufacturer-based formulas for standard setups and focal length ratios for custom designs—especially in OEM and machine vision contexts. Both methods are accurate when used consistently. The key is selecting the model that aligns best with your system’s integration and performance goals.
Important design considerations:
Higher effective magnification reduces the field of view (FoV), potentially requiring more image stitching for large-area inspections.
Increasing magnification without increasing NA does not improve resolution; it just spreads the same resolution over more pixels (oversampling).
System design should align with the Nyquist sampling criterion—at least 2 pixels per smallest resolvable feature are needed to avoid aliasing.
However, this requirement depends on the application:
If the goal is to detect particles (e.g., presence/absence), a 1 µm/pixel resolution may suffice.
If the objective is to inspect shapes, edges, or fine patterns, then Nyquist sampling becomes essential to accurately resolve detail and avoid information loss.
Example:
To image 1 µm features with a 5 MP camera (3.45 µm pixel size), you need at least:
3.45 µm / 1 µm = 3.45 => Need at least 6.9x magnification
This ensures each 1 µm feature spans at least 6.9 µm on the sensor—satisfying the 2-pixel rule.
Resolution and numerical aperture (NA)
While magnification controls how large a feature appears on the sensor, it is the numerical aperture (NA) that governs the system’s ability to resolve fine details. This section explores how NA influences resolution, and how it can be optimized for different imaging scenarios.
Object-side vs. image-side numerical aperture
A complete imaging system includes both object-side and image-side numerical apertures, each affecting different image qualities.
Summary
In microscopic imaging, magnification, NA, and resolution must work together. Optimizing just one won't guarantee better image quality.
Magnification controls how big a feature appears on the sensor. It needs to match the camera’s pixel size to avoid oversampling or undersampling.
NA (Numerical Aperture) determines how small a feature the lens can resolve. Higher NA improves detail and brightness but reduces depth of field.
NA′ (Image-side NA) affects how clearly details are transferred to the sensor. It influences contrast and visibility, especially at high spatial frequencies.
Key takeaway:
High magnification needs high NA for resolution — and both need enough NA′ to maintain contrast. A well-balanced system considers resolution, contrast, and field of view together.
Further reading:
For a detailed exploration of physical and system-level limits, including the Abbe diffraction limit, Rayleigh criterion, optical aberration and Nyquist sampling requirements, continue to "What Really Limits Microscopy Resolution".
To the Article: What Really Limits Microscopy Resolution