An imaging spectroscopic technique is undergoing development for use in remote detection of ice and mapping of the thickness of ice on aircraft surfaces. The technique is based on the variation of local spectral reflectance with the depth of ice and/or water on a surface of known aircraft material (typically, aluminum). The spectrum of white light reflected from each surface point includes absorption dips characteristic of any water or and/or ice present at that point, as distinguished from the relatively flat spectral reflectance of aluminum. Thus, the local thickness of ice (and, optionally, water) can be computed from the local spectral reflectance, and the thickness of ice can be mapped by performing this computation for all points in the image.

Figure 1. The CCD Camera Looking Through the LCTF acquires images of the same surface in 21 adjacent wavelength bands.

In experiments to demonstrate the technique, a band-pass liquid-crystal tunable filter (LCTF) and a 16-bit charge-coupled-device (CCD) camera (see Figure 1) were used to image chilled aluminum cells that were, variously, empty or filled with ice or water to various thicknesses. Reference images of a 99-percent-reflectance standard were also acquired. The aluminum, water, ice, and reflectance-standard images were acquired in 21 wavelength bands, each about 10 nm wide, at nominal pass wavelengths from 850 to 1,050 nm. An independent set of data for verification of the spectral images was acquired by use of a point spectrometer.

Figure 2. The Spectral Reflectance Ratio is defined here as the spectral reflectance of aluminum covered by water or ice ÷ the spectral reflectance of bare aluminum. The spectral reflectance ratio as a function of wavelength can be analyzed to determine the thickness of ice and/or water. Two depths of ice and water, 1 mm and 1.5 mm, were used.

The spectral image data were corrected for CCD dark current and bias and converted to reflectance units, and regions of interest were chosen for determining the spatially averaged reflectance spectra. Some of the results are plotted in Figure 2, which illustrates how spectra can be used to distinguish between, and estimate thicknesses of, water and ice. The experiments revealed one shortcoming; namely, that specular reflection from the surface of interest can cause saturation in affected CCD pixels. Fortunately, saturated pixels can simply be excluded from processing of image data; this was done during the processing of image data in the experiments.

This work was done by Gregory Bearman, Abhijit Biswas, Thomas Chrien, Robert O. Green, and Peter Green of Caltech for NASA's Jet Propulsion Laboratory. NPO-19929



This Brief includes a Technical Support Package (TSP).
Document cover
Optical remote detection of ice on aircraft surfaces

(reference NPO19929) is currently available for download from the TSP library.

Don't have an account? Sign up here.