On the Application of Nighttime Sensors for Rapid Detection of Areas Impacted by Disasters

Abstract

Today a few sensors operating at night are available in the visible/near infrared part of the spectrum, e.g., the U.S. Defense Meteorological Satellite Program Operational Linescan System (DMSP/OLS). However, in the case of DMSP/OLS, the availability of a series of satellites arranged in a constellation and the width of the sensor’s swath allows Earth coverage twice nightly. This can result useful in the aftermath of a natural disaster such as earthquake, when first responders providing relief action need to know the location and the extent of the areas of damages, the potential amount of population involved and the place where survivors are concentrated. Naturally, after this prompt detection of the areas affected by the event, the corresponding very high spatial resolution satellite images can be acquired to obtain an accurate overview of the actual damages. In fact, the availability of a preliminary fast estimate of the areas mainly impacted can support a suitable selection of the very high spatial resolution (VHSR) satellite images acquisition time because these sensors are characterized by a very small frame size that makes unpractical a blind acquisition of the whole region possibly impacted. This way to proceed is also compatible with the longer time usually needed to obtain a VHSR image of a given area of interest, due to the orbital and observation geometry constraints.

Even if it is high, the OLS sensor’s sensitivity could be insufficient to detect settlements with reduced artificial lights, as is often the case in the selected regions of interest. Moreover, in many cases, as for informal settlements following the occurrence of natural or man-made disasters, only the presence of bone-fires could reveal the presence of a human community. As a consequence, it would be necessary to observe the affected areas using wavelengths in the Middle-Wave Infrared region of the spectrum (∼ 4 µm), which is presently not feasible due to the limited sensitivity of available sensors. Nevertheless a couple of examples of the results obtainable using night-time images in these scenarios are provided.

To overcome the difficulties described above, this study focuses on a design analysis of a new night-time sensor. The study is based on accurate simulations of the expected radiance scenario reaching the sensor. This, in fact, is required to assess the characteristics of a new sensor capable of detecting the desired target sources (lights/bone-fires).