SATELLITE SEA-FOG DETECTION SYSTEM USING THE GAUSSIAN MIXTURE MODE, AND FOG DETECTION METHOD WITH UNSUPERVISED LEARNING MODE BASED ON GAUSSIAN MIXTURE MODEL

摘要

The present invention relates to a satellite sea-fog detection system using a Gaussian mixture model and a fog detection method with unsupervised learning mode based on the Gaussian mixture model. In a first aspect, the satellite sea-fog detection system comprises: a clustering module (11) for generating the Gaussian mixture model by performing daily clustering with an EM algorithm; a fog mode determination module (12) for determining a fog mode among clustered elements; and a cloud masking module (14) for performing a removal process using a cloud masking pixel of a COMS among fog pixels and performing a COMS cloud masking test. In a second aspect, the fog detection method with the unsupervised learning mode based on the Gaussian mixture model comprises: a first step of generating the Gaussian mixture model by performing daily clustering with the EM algorithm; a second step of determining the fog mode among the clustered elements; and a third step of performing the removal process using the cloud masking pixel of the COMS among the fog pixels and performing the COMS cloud masking test.