All spectral movies from the flights are run with the “yaPlaySpecX” software, to convert the movies into spectral images of different wavelengths.
To analyse and classify objects from the images, the software “ImageCalculator” is used.
A panorama made from images taken with the SLR camera, during run number 6 with the helicopter over Longyearbyen (the chosen run for analysis). Note that although the central part of the image appears to be free from the effects of stitching the images together, it is visible along the top and bottom edges. The stitching was created with AutoStitch.
Figure 6.1.1: Panorama produced by images taken with the SLR camera.
From this run, 16 spectral images from the whole interval (370—700 nm) are obtained, by choosing a band pass of 22 nm.
Figure 6.2.2 shows all spectral images in our 16 wavelengths of the target line. The difference in signal to noise ratio between the different wavelengths is evident from this figure.
Figure 6.1.2: Image of the target line, in all wavelengths.
This image was used for modelling of the various classes in ImageCalculator.
RGB composite of the target line
A RGB-composite of the run is presented in Figure 6.1.1. The composite is built up by three spectral images, in order to represent the real colours of the target. The three first panels of Figure 6.1.1 shows these three images, captured in wavelengths 634 nm (red), 568 nm (green) and 480 nm (blue). The software yaSpecColorX was used for the first three images, and the composite was produced in ImageCalculator. Even though the real colours aren’t fully representative the RGB clearly shows the colour features of the target line, such as the red roof of the Lompensentret.
Figure 6.1.3: Target viewed in the three different wavelengths (panel 1—3) making up the colour composite (panel 4).
As it is seen on Figure 6.1.3 is the helicopter a very unstable platform for imaging spectroscopy. This means that features which normally are straight looks meandering, such as the road going through Longyearbyen does on Figure 6.1.3. By using the data collected from the Gyro, under the flight, the image could be restored to look like the panorama picture from Figure 6.1.1. This correction is out of scope for this years campaign.