基于不变矩和支持向量机理论的舰船目标识别

Based on moment invariants and the theory of support vector machine ship target recognition

本文利用小波不变矩和支持向量机理论进行舰船目标识别。在识别过程中,首先获取图像,同时将得到的图像转换到极坐标内;然后利用小波矩进行特征提取,在此过程中选择合适的核函数;最后利用支持向量机进行分类。实验结果表明,本文算法具有较高的识别率。

In this paper, the wavelet moment invariants and support vector machine theory were used to identify the ship targets. In the process of recognition, the image was acquired and the image was converted into polar coordinates. Then,the wavelet moments were used to extract the features of the kernel. Finally, SVM was used to classify. The experimental results showed that the proposed algorithm had a high recognition rate.