Abstract:In order to enhance the description ability of features for image target,it is important to fuse different types of features.In this paper,the feature fusion method based on region moment is discussed in detail for ship image target recognition.Region moment is not to compute the moments of image pixels,but image features.Concretely,image features are designed firstly,then the central moment,Hu’s invariant moment and radial moment are used to fuse these features,finally target recognition is realized by KNN classifier.The experimental results show that,compared with region covariance,the region moment method can fuse target features in the visible image or infrared image better and has not the problem of random initialization,this method has better feasibility and effectiveness.