Understanding and choosing the right imaging lens has become more important than ever. Edmund Optics explains some of the key parameters to consider and things to look out for during the selection process.
Imaging and machine vision are becoming more integrated into our daily lives. From autonomous vehicles to advanced medical diagnostics, camera and lens systems are now commonplace. As sensor technology advances, pixel sizes on cameras have been getting smaller, and sensors have grown larger. In the past, cameras were once the limiting component for an imaging system’s performance. Today, lenses have become a critical component in many applications. Even the latest smartphones have had to overcome this limitation by implementing a multiple lens solution.
Resolution is the specification that gets the most attention as megapixels are increasing and pixel sizes are getting smaller. It is helpful to understand how resolution is specified and tested. Simply put, resolution is the smallest feature size on an object an imaging system can resolve. Resolution is usually specified in line-pairs per millimetre (lp/mm) at a contrast percentage. The resolution of a lens is its ability to reproduce object detail onto a sensor. The more line-pairs in a millimetre a lens can resolve, the smaller the resolvable feature size, and the higher the resolution of the lens. Contrast can be thought of as the “crispness” or “sharpness” of the line-pair or feature, and it’s usually specified as a percentage. The modulation transfer function (MTF) curve is a mathematical representation of how a lens reproduces contrast as spatial frequency (or lp/mm) varies.
Resolution can be affected by other specifications such as f/# and wavelength. F/# increases as the aperture is stopped down, usually by closing down the iris. A smaller f/# typically means higher resolution since the diffraction limit of the MTF increases. The diffraction limit can be thought of as the highest performance a lens can achieve according to the laws of physics. The logical assumption would always be to use the smallest f/# possible all the time. But as the f/# decreases, lens performance becomes less limited by the diffraction limit and more affected by optical aberrations. At some point, decreasing the f/# will stop improving the resolution and start reducing the resolution as performance becomes dominated by optical aberrations. This point of diminishing return on performance can vary depending on the lens’s optical design. The design f/# of a lens is often not published by lens manufacturers and can significantly impact the performance, size, and cost of the imaging lens.
There are a great number of optical aberrations that lens designers have to contend with. There are many levers that designers and engineers can push and pull to achieve the desired performance. One of these levers is wavelength, and a large contributor to this is chromatic aberration. Axial colour (chromatic focal shift) occurs when different wavelengths of light have different focus positions along the optical axis. Lateral colour occurs when different wavelengths of light focus to different points on the sensor, causing the image to appear blurry. This effect occurs irrespective of whether a colour or monochrome sensor is used. Whenever possible, using one wavelength of light is best, as monochromatic systems eliminate the issues of chromatic aberrations. In addition, the specific wavelength of light chosen can be important. Similar to how f/# can affect MTF, the wavelength can as well. Using shorter wavelengths of light, such as blue instead of red, can provide higher resolution since they diffract less.
Another drawback to smaller f/#s is a reduced depth of field (DoF). DoF refers to the object distance or depth through which an object can still appear in focus. If the object placement can’t be well controlled, like in many barcode scanning or logistics applications, a higher DoF may be necessary. A trade-off is made between resolution, f/#, and DoF in these cases. Understanding how different optical specifications affect each other is important for properly specifying an imaging system.
All of the specifications and performance criteria discussed above have been in reference to nominal design performance. Nominal design performance can differ greatly from the real-world performance. Nominal design performance does not take into account manufacturing and assembly tolerances. It assumes every component in the lens has been manufactured perfectly. Usually, lens suppliers do not publish their real-world performance data. Because it can vary depending on the sensor size, wavelength, working distance, field of view, and many other parameters, it can be difficult to present a complete view of a lens performance for all applications. Understanding that nominal design performance can be different from toleranced design performance, which can be different from tested performance, is important when discussing specification requirements with a lens supplier. It is much easier to supply a well-performing lens one time than it is to supply multiple well-performing lenses consistently each time. Making sure your lens supplier understands how to design for manufacturability and how to test critical specifications ensures a successful project.
Selecting the best lens for your application is not as simple as it may seem. There are many factors to consider and often compromises to make. This article covered some of the main points but is by no means an exhaustive list, to maximise performance, it is always best to consult with an expert.
