SWIR images evaluation for pedestrian detection in clear visibility conditions

Abstract

Pedestrian detection is a difficult task with significant practical interests especially in the field of Intelligent Transportation Systems. While the problem has been thoroughly studied from a visible, mid-wave infrared and far-wave infrared cameras perspective, little attention has been paid to the short-wave infrared spectrum. But although light in the short-wave infrared region is imperceptible to the eye, due to its reflective nature, it interacts with objects similarly to the visible wavelengths. As a result, short-wave images, including both shadows and contrast, are comparable to visible images in resolution and detail but they are not comparable in color, a feature which makes objects recognizable and yields one of the advantages of the short-waves, namely object or individual identification. In this paper we evaluate the pertinence of employing the short-wave bandwidths for the task of detecting pedestrians in clear visibility conditions.