基于主动形状模型的苹果果形分级研究

苹果果形是评价苹果品质的一个重要指标。为提高果形判别的准确度和效率，首次采用主动形状模型(ASM)的方法进行苹果形态分级。根据试验确定苹果轮廓特征点数为36时为最佳特征点数，然后对不同形状的苹果进行计算机自动标定、校准，运用主成分分析法获取不同形状的苹果模型，并将模型与实际苹果进行灰度匹配，提取像素数目比等特征参数，实现苹果分级。应用结果表明，该方法对苹果果形的判别准确度高达95%，且直观性强、鲁棒性好，具有较好的灵活性。

Identification and classification of apple shape based on active shape models

The shape of the apple is an important parameter for appraising its quality. In order to improve the accuracy and efficiency on apple shape identification, this paper introduces a method based on Active Shape Models(ASM) for the first time. According to experiment, the authors selected the number of thirty-six as the best marking points to describe apple boundary, and all apples in training set were marked automatically. In the following, Procrustes Analysis was used to get generalized shape set, and then Principal Component Analysis(PCA) was employed to get different shape models. In the end, the shape models were used to match real apple and compute the characteristic parameter for getting apple grade. The experiment result demonstrates that the accuracy is up to 95% and the method has good visibility, high flexibility and strong robustness.