An image tracking system for occultation detection of spaceborne spectrometer

When the Satellite-borne very high infrared spectrometer performs atmospheric composition analysis, it needs to use the sun as a light source for occultation detection. The stability of the light source is related to the detection accuracy. In general, it is believed that the light distribution within the maximum solar intensity range is the most stable and it is necessary to accurately identify and track this range. However, due to the inhomogeneous atmosphere and the change of atmospheric density, the solar image will have deformation or block, which affects the accuracy of target recognition and tracking. This paper proposes a multi-objective clustering-based recognition method to solve the problem of accurate multi-target identification under the large dynamic light intensity range, and uses the hardware structure of FPGA+DSP to realize the imaging and tracking system.