Application of Gaussian Mixture Model in Identification of Oil Spill on Sea

Aiming at the phenomena of evaporating, emitting, dripping or leaking of offshore oil platform, optical image acquisition device is used to carry out continuous unattended monitoring and effective oil spill identification algorithm is utilized to monitor and identify offshore oil spill. This paper focuses on exploring the study on application of Gaussian mixture model in segmentation and recognition of offshore oil spill image, describes specific algorithm and build an offshore oil spill model by using the expectation-maximization (EM) algorithm and minimum description length (MDL) principle of Gaussian mixture model and combines sequential maximum a posteriori (SMAP) algorithm to segment and identify oil spill image. The research result shows that this method can be used to effectively acquire oil spill information and effectively segment and identify oil spill image.