Hyperspectral Imaging

Home Hyperspectral Imaging

Hyperspectral imaging has been used for many years to study patterns of plant growth via satellite imaging. This technology has been refined in PlantScreenTM Phenotyping Systems to provide hyperspectral image analysis of plants on a pixel by pixel basis in spectral range from 350 to 2200 nm. Using a hyperspectral camera with an image analysis software, plant reflective indices can be visualized across the entire surface of the imaged samples. These indices may be correlated with numerous physiological conditions, as well as the status of the plant or leaf with respect to chlorophyll content, content of other accessory pigments and metabolites.

Our VNIR and SWIR hyperspectral cameras are, built by our European partners (PSI) and designed specifically for plant phenotyping.  They are available as stand-alone units, or integrated into a range of PlantScreenTM hyperspectral imaging systems that include light sources, reference standards and robotics.  Systems are available for sensor-to-plant and plant-to-sensor applications.  The systems may be deployed in the laboratory, growth cabinet, greenhouse or field.

For indoor applications, a lifting frame holds the motorized linear guide with cameras and light source and movement is synchronized with camera scanning rate. A reference calibration Spectralon object is mounted in the box and calibration with the reference object is automated. The height of the imaging array is controlled via software. The user may move the linear guide manually, with camera and light source, into a side view measuring position. Default, or user defined, measuring protocols are available in software.

Using hyperspectral cameras with image analysis software, plant reflective indices can be visualized across the entire surface of the imaged sample(s). These indices may be correlated with numerous physiological conditions, as well as the status of the plant or leaf with respect to content of chlorophyll, accessory pigments, nitrogen etc.

Examples of published Reflective Indices, all measurable with the PlantScreen™ hyperspectral station include, but are not limited to:

  • Normalized Difference Vegetation Index (NDVI) Reference: Rouse et al. (1974) Equation: NDVI = (RNIR – RRED ) / (RNIR + RRED )
  • Simple Ratio Index (SR) Reference: Jordan (1969); Rouse et al. (1974) Equation: SR = RNIR / RRED
  • Modified Chlorophyll Absorption in Reflectance Index (MCARI1) Reference: Haboudane et al. (2004) Equation: MCARI1 = 1.2 * [2.5 * (R790- R670) – 1.3 * (R790- R550)]
  • Optimized Soil-Adjusted Vegetation Index (OSAVI) Reference: Rondeaux et al. (1996) Equation: OSAVI = (1 + 0.16) * (R790- R670) / (R790- R670 + 0.16)
  • Greenness Index (G) Equation: G = R554 / R677
  • Modified Chlorophyll Absorption in Reflectance Index (MCARI) Reference: Daughtry et al. (2000) Equation: MCARI = [(R700- R670) – 0.2 * (R700- R550)] * (R700/ R670)
  • Transformed CAR Index (TCARI) Reference: Haboudane et al. (2002) Equation: TSARI = 3 * [(R700- R670) – 0.2 * (R700- R550) * (R700/ R670)]
  • Triangular Vegetation Index (TVI) Reference: Broge and Leblanc (2000)Equation: TVI = 0.5 * [120 * (R750- R550) – 200 * (R670- R550)]
  • Zarco-Tejada & Miller Index (ZMI) Reference: Zarco-Tejada et al. (2001) Equation: ZMI = R750 / R710
  • Simple Ratio Pigment Index (SRPI) Reference: Peñuelas et al. (1995) Equation: SRPI = R430 / R680
  • Normalized Phaeophytinization Index (NPQI) Reference: Barnes et al. (1992) Equation: NPQI = (R415- R435) / (R415+ R435)
  • Photochemical Reflectance Index (PRI) Reference: Gamon et al. (1992) Equation: PRI = (R531- R570) / (R531+ R570)
  • Normalized Pigment Chlorophyll Index (NPCI) Reference: Peñuelas et al. (1994) Equation: NPCI = (R680- R430) / (R680+ R430)
  • Carter Indices Reference: Carter (1994), Carter et al. (1996) Equation: Ctr1 = R695 / R420; Ctr2 = R695 / R760
  • Lichtenthaler Indices Reference: Lichtenthaler et al. (1996) Equation: Lic1 = (R790 – R680) / (R790 + R680); Lic2 = R440 / R690
  • Structure Intensive Pigment Index (SIPI) Reference: Peñuelas et al. (1995) Equation: SIPI = (R790- R450) / (R790+ R650)
  • Gitelson and Merzlyak Indices Reference: Gitelson & Merzlyak (1997) Equation: GM1 = R750/ R550; GM2 = R750/ R700)
False color image of NDVI vegetative reflectance index. Chlorophyll degradation dynamics in Arabidopsis Col-0 plants upon herbicide treatment analysed with VNIR hyperspectral imaging.

The PlantScreenTM hyperspectral imaging station allows the user to acquire a full spectral scan across the entire spectral range of the camera for each pixel of the image. The user may select specific wavelengths of interest to record reflective indices that may be correlated with, for example, leaf nitrogen status or the production of anthocyanin to protect photosystem II under high light stresses. It is also possible, in the software, to establish patterns within hyperspectral measurements that are indicative of abiotic and biotic stresses, so that novel protocols for automated stress screening may be established.

For accurate measurements of pigment levels, and various reflective indices, the spectral resolution of the VNIR camera must be at 1 nm or finer. Our VNIR camera has a spectral resolution of 0.8nm with 1920 spectral bands and 1000 spatial bands. The SWIR camera has a range up to 2200 and incorporates the latest hyperspectral chip technology. Cameras from other manufacturers have been shown to have a very limited lifespan with rapid pixel failure.  The PlantScreenTM cameras have overcome this limitation.

VNIR900

SWIR 1700

SWIR 2200