Abstract:

An adaptive mage fire detection algorithm is presented.The kernel function is chosen with the effect of sample data in carful consideration,and corrected by the conformal transformation method of information geometry.Then the corrected function is used to train the SVM to identify fire.The experiment results show that this algorithm improves the classification precision of fire image and accuracy of fire reorganization.

Key words: adaptive;conformal transformation;data driving;fire detection;support vector machine