Recognition of Shaft Orbit Based on Walsh Spectrum and Support Vector Machine

In order to recognize shaft orbits exactly and quickly, a novel recognition method based on Walsh spectrum and support vector machine is proposed in this paper. Firstly, distance vector between the point on the shaft orbit and its center is calculated and obtained. Secondly, the distance vector is transformed by Walsh orthogonal matrix, and obtained Walsh spectrum has property of invariance to rotation, scaling and translation. Thirdly, Walsh spectrum, as feature of shaft orbit, is trained by support vector machine, and classifying functions are obtained. Finally, the remaining samples are tested by the classifying functions, and experimental results are encouraging which demonstrates the effectiveness and robustness of the proposed approach.