New applications and sensors often drive innovation in optics and lighting within the field of machine vision, and this innovation usually must occur simultaneously. The choice of sensors depends largely on the end application in question, and so often on lighting.
For optics, sensors and illumination are often “problematic applications,” requiring solutions to challenging problems. Even highly sensitive sensors sometimes fail to function unless light is correctly and evenly distributed within the lens’s effective area. The challenge for optics suppliers is keeping pace with the rapid evolution of machine vision applications.
With the advancements in sensor and lighting technologies, the demand for new applications is so vast that the question facing optical product suppliers has become, "Which lens designs need to be optimized to work in tandem with the applications that are emerging downstream?" says Greg Hollows, Vice President of Imaging at Edmund Optics.
Challenges of optical design
Remote sensing devices mounted on drones may prefer large-format sensors with high resolution, simply because more pixels can be packed into a larger array, allowing the drone to collect as much image data as possible within a limited flight time. This application requirement is easy to understand if you are a sensor manufacturer. Therefore, actively investing in the development of larger sensors makes sense. However, for optics manufacturers, the choice is not so simple, as they cannot design larger lenses without considering other factors.
Many systems combine lens components and filters together, which can become even more complex for wide-field-of-view applications that are often associated with drones today.
Co-design with lighting
This demand for a holistic design approach extends to lighting components. Ideally, the lighting and imaging optics should be designed as a unified system to ensure that the lens collects the best light, not the suboptimal. Co-designing lighting and optics makes it easier to match the two for optimal results. It also minimizes costs and shortens time to market.
The wavelength range of the light source in an imaging system also affects lens selection and overall system performance. Narrowing the range to a few nanometers typically improves image contrast. Using filters with broadband light sources can be helpful, but such solutions are not always as flexible as narrowband light sources.
The wavelength of a light source can be used to accomplish many things in machine vision . Using narrowband light sources such as LEDs minimizes the corrections required for lens design and introduces interesting new options for increasing image contrast. The way specific materials absorb or reflect specific wavelengths can also highlight defects that are invisible under broadband light sources.
Within the visible light range, every color has an opposite color that can be used to enhance image contrast. Therefore, illuminating a green object with a green light source will make it appear brighter on the image sensor, while illuminating it with red light (the opposite color of green) will make it appear darker.
However, improving image contrast doesn't necessarily require narrow-band light sources. For example, shorter wavelengths in the ultraviolet (UV) range scatter surface features more strongly than visible or infrared light. Alternatively, if UV light is absorbed, it tends to be absorbed at surfaces. In either case, using gas discharge tubes or UVLEDs to image how UV light interacts with materials can help detect contaminants or shallow scratches that are invisible in other regions of the spectrum. This benefit is particularly evident when examining glass displays, lenses, and other materials that appear transparent in the visible spectrum.
Disclaimer: This article is a reprint. If it involves copyright issues, please contact us promptly for deletion (QQ: 2737591964). We apologize for any inconvenience.
