Comparison between skylight polarization in visible and ultraviolet spectra for heading measurements

• Moutenet Antoine
• Serres Julien
• Viollet Stéphane

• Ultraviolet UV
• Optical compass
• Skylight polarization
• Rayleigh scattering
• Celestial compass
• Polarized vision

In the field of autonomous outdoor navigation, the calculation of a precise heading is essential. For more than one century, numerous studies have been carried out on the ability of navigating insects, such as desert ants Cataglyphis fortis, to find their bearings thanks to the polarization of light scattered by the atmosphere. In space, the sunlight is naturally unpolarized, however, it is scattered by atmospheric molecules, and then becomes polarized according to a very precise pattern seen from the ground. This makes it possible to find the sun’s position, even if the sun is hidden. Although it has already been possible to build celestial compasses based on this polarization phenomenon, it tends to be corrupted in the visible spectrum on overcast days, limiting their use. However, still relying on insects which initially inspired us, we can see that they are sensitive not only to visible light polarization, but also to ultraviolet polarization, which could be the key to navigating under overcast skies using polarimetry. We therefore decided to setup a UV polarimetric camera (UV-sensitive camera with a rotating UV linear polarizer and an objective directly mounted onto it, composed of a UV-centered band-pass filter and a pinhole of 50μm or 200μm), because it was not commercially available, and to compare its levels of performance with a conventional visible polarimetric camera (polarized monochromatic camera), under different weather conditions. Our preliminary results with our UV-sensitive camera in overcast conditions are not yet as good as those we get from our visible camera in clear conditions, notably because of the very long integration time intrinsic to the UV-sensitive sensor’s rusticity. Despite mixed results, this study does demonstrate the advantage we can get by looking into this new type of sensors.