Research On Red Tide Detection Technology Of MODIS Image Based On Association Classification Method

In recent years, China’s coastal Red tide disasters occur frequently and have been expanding in scale, leading to unbalance and deterioration of the marine ecosystem, and bringing serious damage to marine economy. The technology of the red tide detecting is pivotal to prevent and reduce the outbreak number of the red tide disaster. Because the restriction of conventional sea surveys, the remote sensing technology has been more and more applied in the study on detection of the red tide. Red Tide detection can be regarded as a classification of red tide and seawater as two classes. The Associative classification methods not only can solve the classification problem, but also can reflect the relationship of a variety of factors related to red tide, with which can reflects the advantages in red tide detection based on remote sensing images. So this paper studies the red tide detection method of MODIS images based on Associative classification methods. The main work in this paper is as follows:In this paper, the associative classification method is used to mining the classification rules which can be used to distinguish the red tide from the sea water in MODIS images. And then, the red tide detection rules are applied to detect the red tide, the result showing that the red tide detection method based on classification associated rules can detect the area where the red tide outbreak.For that single red tide detection rules can’t detect the multi-algal red tides effectively, a multi-algal red tide detection method is proposed. The associative classification method is applied to mining the rules to detect the multi-algae red tide. The result of the experiment shows that the rules not only ensure the efficiency and the accuracy of detection of red tide, but also can determine the dominant species from the red tide types that have been detected.