The Research On MODIS Data Processing And Marine Oil Spill Classification

With the blooming of the economy, oil spill accidents happen frequently, which cause immeasurable economic losses and catastrophic disaster about ecological environment. Because of the complexity of ocean situation, the shape of oil spill area changes fast, it is great difficult and cost-consuming to eliminate the damage.So it is very important and meaningful to make use of the various ways about remote sensing technology to carry on the early warning and real-time monitoring of ocean oil spill. As the main and important detection instrument onboard the TERRA and AQUA satellites, MODIS has more advantages and characteristics than NOAA-AVHRR and other Landsats, besides the higher spatial resolution, more frequently it transits and higher spectral resolution, it also monitors more extensive area with lower cost (National Aeronautics and Space Administration (NASA) has policy that it’s free to receive the MODIS data for global). These characteristics are helpful to enhance the ability of observing, recognizing and real-time monitoring the type of surface. Now, MODIS data has been widely used in the fields of monitoring ocean oil spill events (like red tide, seaice, ocean oil spill etc) and so on, it has value of important theoretical study and engineering application.MODIS is multi-spectral data (general 36 bands), Based on the MODIS data analysis and the spectrum calibration of overlapped areas, this paper, which is part of China 863 project, focuses on the study of adaptive band selection method, then constructs the oil spill area feature vector with the method, and studies the problem of classification of ocean oil spill area. Generally speaking, The main work is listed below:(1) The building of the MODIS Process and Analysis System. MODIS’s memory type is Hierarchical Data Format (HDF), with the library of HDF offered by HDF-group attached to NASA, this paper finishes the building of system, which has functions that MODIS data analysis, pre-processing like the elimination of overlapped area and adaptive band selection. Now, this system is operated by National ocean monitoring forecasting center (Qingdao) at trial.(2) Carrying out and finishing the classification of oil spill area on MODIS data image. With the adaptive band selection on built system, feature vector on oil spill area is formed, and oil spill area is classified and marked based on the feature vector, the probable result of oil spill classification is given. The experiment result shows after the classification of the 2011 Bohai Gulf and 2010 Mexico Bay oil spill data, this method shows high execution efficiency, effective ability of classifying the oil spill area and the value of engineering application study. Due to the limit of test data, its accuracy of classification is certain distance away from business application, it should be further improved and enhanced.