基于Zernike矩的水果形状分类

Fruit shape classification based on Zernike moments

提出了一种基于Zernike矩的水果形状分类方法,首先运用标准矩对水果图像进行归一化,使得归一化后的图像具有平移和尺度不变性,然后从归一化后的图像中提取具有旋转不变性的Zernike矩特征,并运用主成分分析法确定分类需要的特征数目,最后将这些特征输入到支持向量机分类器中,完成水果形状的分类.通过与傅立叶描述子的分类性能比较,结果表明由于Zernike矩具有良好的正交性和旋转不变性,使分类性能明显有大幅提高.

An approach for fruit shape classification based on Zernike moments was proposed. The image was first subjected to a normalization process by using its regular moments to obtain scale and franslation invariance. The rotation invariant Zernike features were then extracted from the scale and translation normalized images and the numbers of features were decided by primary component analysis （ PCA ）. At last, these features were inputted to support vector machine （SVM） classifier to finish the shape classification. This method works better than traditional approaches because of the orthogonal base and rotation invariance of the defined features, which was verified by experiments on Zeruike moments and Fourier descriptors.