An Infrared Hot Spot Recognition System On-Board of the Small Satellite BIRD

Abstract

The observation of high temperature events as forest, vegetation and coal seam fires, volcanic activity or burning oil wells is an interesting part of global monitoring and is important for the estimation of the impact on environment, atmosphere, and climate. The BIRD (Bi-spectral Infrared Detection) to be launched in October 2001 is the first satellite which is equipped with space instrumentation dedicated to recognize high temperature events. Current remote sensing systems being used for the observation of these events have significant drawbacks (low temperature channel saturation, low spatial resolution) because they are not designed for the hot spot investigation. The German BIRD demonstrator mission shall answer technological and scientific questions related to the operation of a compact bi-spectral infrared push-broom sensor on board of a micro satellite and related to the detection and investigation of fires from space. This report is discussing the basic ideas for fire detection and the estimation of fire temperature and fire size in the sub-pixel domain and describes the technical solution for the infrared sensor system on board of BIRD. Results of airborne-based measurements during forest field fire experiment show the ability of the system to solve this task. The payload of BIRD is a multi-sensor system designed to fulfil the scientific requirements under the conditions of a micro satellite. A quantitative analysis of high temperature events from space requires a system which is able to measure up to 1200K without saturation but also detect fires with an area extension of some ten square meters. This requires a large dynamic range that has been realised by a special sampling method within a push broom scanner technique. Besides measurements of normal thermal scenes in the two IR channels, an internal signal processing selects hot measurements. A quantitative determination of such fire parameters as fire temperature and burning area can be retrieved by a bi-spectral method (Dozier 1981), using a Midwave IR channel at 3 - 5 µm (MIR) and a Thermal IR channel at 8 – 10 µm (TIR). The multi-sensor system consists of the following main parts: ·an infrared array sensor system based on cooled CdHgTe-detectors with 2x512 pixels in a medium infrared channel (3.4-4.2µm) and in a thermal infrared channel (8.5-9.3µm) ·a CCD camera with a red and a near infrared channel ·a payload data handling system with a 1 Gbit mass memory ·an on-board neural network classificator. The simultaneous co-registration of VIS and Infrared channels is a key feature of the BIRD equipment for the recognition of high temperature events. The results of the fire experiments under leadership of the group of Fire Ecology of the University of Freiburg with the “Advanced BIRD Airborne Simulator” in September 2001 with a BIRD like equipment on-board on an aircraft have shown that the estimation of fire temperature based on the Doziers method taken by the Infrared sensors fulfil the expectations to these experiments. First results of BIRD measurements of the remote sensing of forest fires will be presented.