How do you think the new GigE standards will influence the machine vision industry?
Respond or ask your question now!
Multispectral imaging involves capturing images of a scene or object over multiple discrete wavelength bands and extracting spectral content from that data. By leveraging known spectral absorption or emission features to identify materials, the technique can be used for everything from mapping rock types in geological formations to identifying blood oxygenation or cancer cells in a laboratory. The problem is that multispectral imagers have historically been large, expensive, sophisticated airborne or satellite-mounted instruments. Because each scene is captured in three-dimensions (x, y, λ), the resultant data cubes are gigabytes in size, only a fraction of which is useful data. Even though multispectral imaging would be a beneficial tool for a range of low-cost, real-time, limited-wavelength applications like anticounterfeiting measures or medical diagnostics, the logistics make it impossible. Our new lithographically patterned dichroic approach presents a way to change that.
A typical multispectral imager essentially consists of either rotating filter wheels or mechanically diced thin-film dichroic filters mounted in front of an image sensor. A hyperspectral imager consists of a dispersing element coupled to an image sensor. The most common applications call for these imagers to be flown over the landscape by airplane or satellite to capture images of the swath of land below. Designs based on line-array detectors leverage cross-track scanning systems to image over the x-axis while they are carried along the y-axis (whisk broom), while those using 2D area detectors capture the full field of view without scanning (push broom). Dispersing elements such as slits, gratings, liquid crystals, or acousto-optic tunable filters allow the system to image the scene at dozens to hundreds of spectral channels, each only nanometers wide.
Such multispectral systems tend to be research-based and can cost hundreds of thousands, or even millions, of dollars. Even for those touted as commercial systems, there is no real volume production pathway to as few as thousands of units. In addition, these systems suffer from being not being able to simultaneously capture the entire scene and all of the spectral content at once. Reaching that goal requires first rethinking the problem.
Scientific studies like satellite mapping may need hundreds of spectral channels, but what about applications that require data over only a few well-defined wavelengths of interest? In our experience, this is a common syndrome: A researcher views his or her process at 100 distinct wavelengths, but in the end, only three to five wavelengths show anything of interest. The applications of this group present dramatically lower performance requirements, which allow us to offer an economical alternative. Using our lithographically patterned dichroic filters, we can build multispectral imagers that capture a scene over a small number of spectral channels. More important, such filters can be fabricated quickly, economically, and reliably using well-established high-volume batch-processing techniques.
The Bayer Filter
Digital cameras typically capture color images using an approach called Bayer filtering.1 In a Bayer filter mosaic, each quartet of pixels consists of one pixel filtered to transmit only red light, one pixel that transmits only blue light, and two that transmit only green (see Figure 1). This filter overlays the image sensor pixels with one-to-one correlation; in other words, when the detector captures an image, 25 percent of the pixels capture red wavelengths, 25 percent capture blue wavelengths, and 50 percent capture green wavelengths. These RGB filters are typically absorptive gels and have extremely wide absorption bands with significant spectral cross talk between the red, green and blue channels. The resultant data is processed using color-space interpolation algorithms to create a color image.
We can apply this same architecture to our patterned multilayer dichroic optical filters. We begin with the Bayer filter concept but instead of using broad absorptive gels to capture data at red, green, and blue wavelengths, we custom design the dichroic filters so that the image sensor operates at λ1, λ2, λ3, and so on.
Dichroic optical coatings are interference filters designed to have different spectral characteristics, either reflection or transmission (i.e. antireflection or bandpass). Eight years ago, Ocean Optics developed the technology of combining deposition with lithographic patterning techniques to fabricate structured dichroic optical filter coatings.2,3 Using this methodology and refining it over time, we have produced hundreds of thousands of sharp-edged optical filter structures for a variety of applications, including spectral sensing and imaging, LCD and DLP displays, and entertainment lighting.
The coatings consist of multilayer stacks of high- and low-index materials fabricated via plasma-assisted deposition or sputtering. Adjusting the layer thickness and number of pairs tunes the spectral characteristics of the resultant coating to reflect or transmit over specific bands. The lithographic step allows us to pattern an entire wafer in pixels as small as 10 Ám on a side.
By alternating between deposition and lithographic patterning steps, we can form Bayer-filter-like structures corresponding to the wavelengths of interest for a specific application. We deposit the layers for the λ1 filter, lift off the patterning, re-pattern the wafer to apply the λ2 filter, and so on. In experiments, we have iterated through 11 successive depositions and patterns in a matter of a few days. The technology allows us to make a custom multispectral imager based on what the customer needs, with a direct pathway to high-volume production (see Figure 2).