To realize the effective recognition of pebrine image, on the basis of the sample preparation and shape feature analysis of pebrine image, this paper focuses on the research of multifeature parameters extraction technology of pebrine image. In particular, in order to realize the direct separation of colorized pebrine image from complex background, the method of color feature extraction criterion for pebrine image has been proposed. In addition, the research on the multi-feature parameters extraction technology of the initial shape feature sets has been done to determine the optimal feature set of classification for pebrine image, which adopts the optimal feature selection method based on the learning of the classifier. The result shows that the best classified feature set has good effect of classification, which is formed by the combination of geometric shape features and Hu's invariant moments feature, and the method can decrease the dimensions of the features, as well as increase the accuracy of classification.