Image Target Recognition Algorithm Based On The Msea And The Same Moment

This paper focuses on the research of image feature matching based on moment invariants and the study of template matching algorithm based on MSEA. According to the basic principle of image recognition. MSEA (Multilevel successive elimination algorithm) is showed to reform the original template matching algorithm. It is a fast ex~haustive search algorithm and simulation results demonstrate that although the performance of the proposed algorithm is the same as that using the exhaustive search, the computation time has been reduced significantly. Aim at the targets which has been translated~ scaled and rotated, the paper presents the method which uses moment invariants as the feature and the comprehensive study of the effectiveness of different moment invariants such as Bamieh., Zernike and Hu抯 moment in pattern recognition applications, the experimental results show that these algorithms are effective and well used. Moreover, region segment method combined projection and 搑eptile searching?is used to pick-up the region existed the single undiscriminating targets after the local max of wavelet transform of noise images. In order to ensure veracity of the recognition results, 50 samples of 5 classes were trained for testing. The classification procedure was based on the k- nearest neighbor rule and the multiple membership classification. The comparing result of each moment invariants shows that all the moments have good performance of recognition and I-IIMI is easy computing and steady. Based on HIvll, the ratio of scale and rotation are computed by the principal axes and the centroid of the targets.