Spectral imaging system

Details Spectral imaging system Spectral imaging system

2000-08-07

2002-04-16

Evans, F. L. (Department: 2877)

Optics: measuring and testing

By dispersed light spectroscopy

Utilizing a spectrometer

C356S328000

Reexamination Certificate

active

06373568

ABSTRACT:

BACKGROUND OF THE INVENTION
In this application, a spectral weighting function that indicates the presence or amount of a certain trait in a sample is termed a spectral index, spectral weighted index or spectral measure for that trait. The term spectral measure is also sometimes used to denote the measurement value obtained at a given point or region according to a given spectral weighting function.
Various spectral imaging systems are used to derive spectral information about samples, including imaging spectrometers, imaging interferometers, band-sequential cameras using e.g. filter wheels, and linear-variable filter imagers. Those systems which are based on linear-variable filters or on spectrometers generally acquire data about a single point or line at a time, and optical or mechanical means are used to develop a two-dimensional image of the sample. In many cases, a fully-populated image cube is obtained, consisting of an image of the sample at each adjacent spectral band in the spectral region of interest. With interferometers, linear-variable filters and most spectrometers, there is no way to avoid acquiring the full image cube, even if only a portion of the cube is desired. Data requirements are enormous: for high-resolution images (2048×3072 pixels, 12 bits/pixel) with 32 spectral bands, nearly 400 megabytes of data must be acquired and processed. Filter wheels enable selective acquisition of the wavelengths of interest, so a sparsely populated cube may be obtained. However, if there are N bands of interest, a minimum of N exposures is still required. This is quite time-consuming and inefficient.
Spectral measures have been developed that utilize many spectral bands to describe a sample attribute, or to classify regions within a sample. An example is described in “Biomedical Applications Of The Information-Efficient Spectral Imaging Sensor (ISIS)”, Gentry, Steven M.; Levenson, Richard M. Proc. SPIE Vol. 3603, p. 129-142 (1999), which describes a technique called projection pursuit that relies on a specific weighting of various spectral bands to derive an enhanced detection of sample properties. Other well-known methods such as principal component analysis involve generating spectrally-weighted sums with precise control over the weighting function.
Gentry also describes an imaging system which he terms an information-efficient spectral imaging system (ISIS). ISIS provides means for imaging a line of the sample at a time through a dispersive element, followed by an element that selectively transmits light in each spectral band according to a desired transmission amount, then a detector which determines the spectrally-weighted signal content at each pixel along the line. This produces an optically-weighted signal at the detector, rather than requiring digitization of each spectral component followed by numerical integration. While efficient in that sense, which is a great advance over the prior art of interferometry or band-sequential filters, it only images a line at a time and so requires several thousand exposures to produce a two-dimensional image. Stepping or scanning means must be provided as well, to sweep across the sample.
Use of lamps and mechanically-tuned gratings to produce monochromatic light, and thus to obtain spectral images, is well-known in the art.
The use of red, green and blue LEDs for illumination is known in the art, both in time-sequential fashion or simultaneously. The colors produced by such systems are broad, ill-defined, and would not be suitable for applications that involve precise quantitative spectral measures or indicia.
For example, in U.S. Pat. No. 5,838,451, McCarthy teaches the use of a handful of LEDs with broad spectral outputs (40 nm or more) to illuminate a sample, either singly or in combination to approximate standard functions such as the tristimulus curves. These are used to construct a calorimeter. To ensure stability in time, the temperature of the LEDS is precisely regulated. Such a system is unsuitable for spectroscopic measurements, nor can it obtain accurate images of spatial distributions within objects because the illumination patterns of the various LEDs are not spatially alike.
McCarthy teaches in U.S. Pat. No. 5,137,364 a system wherein two or more LEDs with different broad emission spectra are used to illuminate a sample and the resultant light is imaged by two or more photosensors, each of which has a different spectral responsivity. The source LEDs are turned on one at a time in sequence, and the photosensor readings are noted in each case. From the table containing readings from each photosensor under each of the illumination conditions, and a calibration, various spectral and/or colorimetric estimates of the sample are obtained.
U.S. Pat. Nos. 5,608,213 and 5,567,937 describe the use of infrared LEDs as scene simulators to test night-vision goggles with a mixture of wavelengths that approximate terrestrial scenes.
The use of an array of LEDs with a current switch, along with optics including a dispersive element such as a prism or grating, to produce light of time-sequential wavelength which illuminates samples for subsequent spectral analysis is described in U.S. Pat. No. 5,029,245 issued to Keranen.
A subsequent patent, U.S. Pat. No. 6,075,595, uses the same apparatus as Keranen and further incorporates a substrate with embossed dimple-shaped light concentrators in which the LEDs are mounted, to restrict the angles at which light emerges from the LEDs and passes into the collection optics. This is said to enhance brightness and reduce scatter.
Further discussion of LED array illuminators are contained in “Novel Spectroscopic Techniques For Biomedical Applications”, Hyvarinen, T.; Aikio, M.; Esko, H.; Malinen, J.; Proc. SPIE Vol. 2084, p. 224-230 (1994); and “Thirty-Two Channel Led Array Spectrometer Module With Compact Optomechanical Construction”, Malinen, J.; Keranen, H.; Hannula, T.; Hyvarinen, T., in “Optomechanics And Dimensional Stability; Proceedings Of The Meeting”, San Diego, Calif., Jul. 25, 26, 1991, pp. 122-128 (1991).
Ken Spring of the N.I.H. has demonstrated a system that uses three or four LEDs for time-sequential excitation of fluorescent samples at distinct excitation wavelengths. The optical arrangement collimates the emission from each LED and then passes it through a bandpass filter to define a spectral band. The collimated light from the several LEDs is introduced into the pupil plane of a telescope, which accepts light from whichever LED is lit, and feeds it into a multimode fiber. The fiber spatially scrambles the light and removes any inhomogeneities, after which the light emerges from the fiber and is re-imaged to illuminate the fluorescent sample. This system creates light that is essentially monochromatic at any point in time.
There is nowhere disclosed in the prior art a means for imaging a two-dimensional sample so as to directly obtain an image of some spectral index at every point in the sample. Nor does the prior art teach means for interactively determining such a spectral index from a first sample or region of a sample, then imaging other two-dimensional regions or samples according to that spectral index, except through the use of interferometers or band-sequential filters to obtain image cubes or the equivalent, each time an image is desired.
It is an object of the present invention to provide a system and method for obtaining images that reveal the precise value of potentially complicated spectral indices at every point in a two-dimensional image, without need for acquiring a complete image cube. It is another object to provide means for determining these spectral indices in situ, along with means for calibrating or compensating for the spectral response of the optics and detector used in the measurement. It is yet a further object to take advantage of spectral indices derived using techniques such as projection pursuit, principal component analysis, and the like, to distinguish between species or quantify amounts of species within a sample. A further aim is to achi

Affiliated with

Also associated with

Rating

Spectral imaging system does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Spectral imaging system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Spectral imaging system will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2880323