8/6/2019 2006 10 02 Multi Spectral Cameras

http://slidepdf.com/reader/full/2006-10-02-multi-spectral-cameras 1/2

125 Tech Park Drive, Rochester, NY 14623 1-585-427-8310 (voice) 1-585-427-8422 (fax)www.geospatialsystems.com

© 2006 Geospatial Systems, Inc.

Multispectral Cameras

Expanding Beyond the Visible

Modern vision and imaging applications rely on interpretation of information acquired by an image sensor. Typically the sensor isdesigned to emulate human vision, resulting in a color or monochrome image of the field of view as seen by the eye. This isaccomplished by sensing light at wavelengths in the visible spectrum (400-700 nm). However, additional information can be gainedby creating an image based on the light that is outside the sensitivity of the human eye.The information available can be maximized by combining information found in multiplespectral bands. The photonic spectrum includes energy at wavelengths ranging from theultraviolet through the visible, near infrared, far infrared, and finally, x-rays. The color image from a Charge Coupled Device (CCD) array is acquired by sensing thewavelengths corresponding to red, green, and blue light.

CCD sensors are capable of detecting light beyond the visible wavelengths out to 1100nm. The wavelengths from 700 nm to 1100 nm are known as the near infrared (NIR) andare not visible to the eye. In standard color video cameras the infrared light is usually blocked from the CCD sensor because itinterferes with the quality of the visible image.

GSI’s multispectral cameras give you access to the full power of the CCD’s capabilities by providing one imaging array that performscolor imaging and two more that sense the invisible light from 700-1100 nm. The wavelengths detected by each array can be further limited by adding narrowband optical filters in the imaging path. Combining the information from all three sensors provides imagedata in five spectral bands: red, green, blue and the two infrared bands.

GSI’s multispectral cameras can acquire images fromup to five spectral bands. Light enters the camerathrough the lens and is separated by a dichroic prism.The visible wavelengths (400-700 nm) are directed toa color CCD array that images the red, green, and bluebands. Near infrared wavelengths (700-1100 nm) aresplit between two monochrome CCDs. Optionalnarrowband filters can be placed in front of the arraysto select specific wavebands. These trim filters can becustomer specified and custom built at the factory

allowing optimization for your specific application requirements. The color array can be replaced with a monochrome array

providing a custom three-color, three-CCD camera. The prism can be fabricated with user specified dichroic coatings for OEMapplications.

Pixel data from each of the three arrays is digitized and processed by the Digital Signal Processing module. This module isprogrammable and can be customized to meet application requirements. A variety of operations such as comparisons betweenimages or thresholding and overlays can be performed in real time. The image data from each of the three CCDs and anyprocessed images can be selected for display. An encoder synthesizes analog NTSC, PAL, and S-Video formats. An optionaldigital interface allows direct access to the digital pixel data.

Advanced optical design and CCD alignment features result in images from all three channels that are the same size andregistered to within a fraction of a pixel. This image registration and square pixels provide high quality images ideal for advancedvision applications.

Multispectral Imaging

8/6/2019 2006 10 02 Multi Spectral Cameras

http://slidepdf.com/reader/full/2006-10-02-multi-spectral-cameras 2/2

125 Tech Park Drive, Rochester, NY 14623 1-585-427-8310 (voice) 1-585-427-8422 (fax)www.geospatialsystems.com

© 2006 Geospatial Systems, Inc.

Food ProcessingIn applications like produce sorting, the elimination of pigmentation effectsallows easier identification of defects. By combining color and infrared

imaging, a single camera can be used tosort for defects and grade for color appearance or ripeness.

Precision AgriculturePlants have very high spectral reflectivity in the NIR. THis reflectivity is associated withchlorophyll and xanthophyll content. By studying changes in reflectivity, growers can identifyplant stress resulting from problems such as water shortage, nutrient deficiencies, or toxinslong before there is any visible indication.

Environmental Imaging and ReconnaissanceInfrared film has long been used to analyze the healthand composition of vegetation. However, working withfilm requires delays for processing and handling. Bymapping the green, red, and a NIR band to the blue,green, and red channels of the output video signal real-time false color-IR video is created.

The different characteristics of spectral imaging canprovide a number of unique advantages in visionsystems. The longer wavelengths of infrared penetratehaze and fog better than visible wavelengths, resultingin enhanced visibility under these conditions.Differences in spectral reflectivity can distinguishdifferent objects that appear the same in the visible.

Defects such as bruising onfruit are more easily isolated in an infrared image.

A single camera can sort for ripeness and defects.

Multispectral BenefitsThe ability to “see” objects with a camera and detect color is a powerful tool. By processing the data from an image, additionalinformation can be extracted. Multispectral imaging expands the camera’s capability to include the power to image features that cannot be seen with the eye. By selectively combining both visible and infrared images, the available information from a field of view canbe maximized.

These images illustrate the use of color mapping. The plastic bag and shirt are well camouflaged in thevisible image and very obvious inthe infrared. The false color imagefurther distinguishes the vegetation.

Most vegetation shows a largeincrease in reflectivity in the near infrared.

Although this ivy and itssilk imposter look thesame in the visible, thereal plant is easily identified in the infrared by its high reflectivity.

Color

Color

Color

Infrared

Infrared

Infrared

False Color

Color

Infrared