Image analysis – Pattern recognition – Feature extraction
Reexamination Certificate
2001-07-26
2004-11-16
Mehta, Bhavesh M. (Department: 2625)
Image analysis
Pattern recognition
Feature extraction
C382S191000, C382S260000, C250S316100, C250S339090, C073S023370
Reexamination Certificate
active
06819798
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates generally to the field of imaging systems, and more particularly, to hyperspectral imaging systems. The invention further relates to a method for determining the magnitude and direction of spectral drift in hyperspectral imaging systems.
2. Related Art
All objects, such as soil, water, trees, vegetation, structures, metals, paints, fabrics, etc., create a unique spectral fingerprint. The unique spectral fingerprint is a measure of the electromagnetic energy that is reflected or thermally generated from an object when the object is being imaged. A sensor is used to determine the fingerprint. The sensor measures the light reflected from the object or the thermally radiated energy from the object, the majority of which registers in wavelengths or bands, invisible to humans.
Hyperspectral imaging refers to the imaging of an object over a large number of discrete, contiguous spectral bands. Hyperspectral imaging produces images for which a spectral signature is associated with each spatial resolution element or pixel. A cross between imaging and spectroscopy, hyperspectral imaging is an outgrowth of multi-spectral imaging. Multi-spectral imagers take one measurement in a wide portion of each major wavelength band. Unlike multi-spectral imagers, hyperspectral imagers measure energy in numerous narrow units of each band. The narrower bandwidths of a hyperspectral sensor are more sensitive to subtle variations in energy reflection or thermal generation. As a result, the hyperspectral signal is more detailed and contains more specific information about the object imaged.
A hyperspectral imager can see what cannot be seen by the human eye. Hyperspectral imaging, combined with image processing and analysis, can identify materials, classify features, identify trends, etc. Hyperspectral imaging can be used for, but is not limited to, medical photo-diagnosis, chemical detection, cloud tracking, earth resources, remote sensing of pollutants and other compounds in the atmosphere, soil, and waters, remote sensing of land and oceans, and target detection and recognition.
Hyperspectral imaging systems require precise control and identification of spectral channel assignments during image data recording. For precise radiometric and spectral detection algorithms to be effective it is essential to maintain spectral channel calibration during operation. Knowledge of the spectral bands during image processing allows utilization of a wide range of algorithms to determine material identification based upon spectral signature information.
Identification of the spectral channel assignments during imaging operations is required to determine the amount of spectral drift which can occur during mission environments and provide the information necessary to correctly reassign the spectral channels to the hyperspectral imagery. What is needed is a system and method for detecting spectral drift. What is further needed is a system and method for determining the amount as well as the direction of spectral drift.
SUMMARY OF THE INVENTION
The present invention is a method for detecting magnitude and direction of spectral drift in a hyperspectral imaging system. More particularly, the present invention is a method for determining magnitude and direction of spectral channel drift for several contiguous spectral regions over a wide spectral range in hyperspectral imaging systems. According to the method of the present invention, in-field testing with a spectral filter sequentially irradiated by two blackbody sources is performed to generate a response function of the spectral filter. The response function is ensemble averaged to reduce any noise. Background radiance is then removed to produce a smoothed spectral transmittance function of the spectral filter. The first derivative function of the spectral transmittance function is then determined. The first derivative function is partitioned into spectral band regions having +/− N pixels on either side of the function minima. The value of N is selected to optimize the detection algorithm sensitivity to change while extending the limit of spectral shift magnitude. The sum of the differences between the first derivative function and a reference spectral derivative function is determined. The difference result is applied to a look-up table to determine magnitude and direction of spectral drift for each of the partitioned spectral band regions.
Use of the present invention can provide information on spectral distortion or spectral smile for 2-D focal plane arrays used for hyperspectral imaging.
REFERENCES:
patent: 4873481 (1989-10-01), Nelson et al.
patent: 5424543 (1995-06-01), Dombrowski et al.
patent: 5602394 (1997-02-01), Dombrowski et al.
patent: 5822222 (1998-10-01), Kaplinsky et al.
patent: 6008492 (1999-12-01), Slater et al.
patent: 6180990 (2001-01-01), Claiborne et al.
patent: 6341257 (2002-01-01), Haaland
“Infrared Transmission Wavelength Standard”, SRM Catalog [online]. National Institute of Standards & Technology Certificate, [Retrieved Dec. 9, 1999]. Retrieved from the Internet: <URL: http://ois.nist.gov/srmcatalog/certificates/view_cert2.cfm?certificate+1921a>, 7 pages.
Simonds, J.A. et al., “Internal Diagnostics for FT-IR Spectrometry”,SPIE, vol. 3082, pp. 106-120.
Zhu, C. and Hanssen, L.M., “Absorption Line Evaluation Methods for Wavelength Standards”,SPIE Conference on Optical Diagnostic Methods for Inorganic Transmissive Materials, San Diego, California, Jul. 1998, SPIE, vol. 3425, pp. 112-118.
Copy of International Search Report issued Apr. 10, 2002 for Appln. No. PCT/US01/23418, 7 pages.
BAE Systems Information & Electronic Systems Integration, Inc.
Chawan Sheela
Mehta Bhavesh M.
Sterne Kessler Goldstein & Fox P.L.L.C.
LandOfFree
Spectral drift and correction technique for hyperspectral... 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 drift and correction technique for hyperspectral..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Spectral drift and correction technique for hyperspectral... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3317789