摘要
提出了一种基于灰度-梯度共生矩阵的稀疏表示声纳图像识别方法。该方法采用灰度-梯度共生矩阵对声纳图像进行特征提取,特征提取结果相比全局的特征提取包含了声纳图像的重要的纹理信息;并结合稀疏表示的分类方法对声纳图像进行识别。实验表明,该方法既满足了对声纳图像进行识别实时性,又提高了识别的准确性。
This paper proposes a new sonar image recognition method based on gray level-gradient co-occurrence matrix sparse representation.The gray level-gradient co-occurrence matrix is used to extract the feature of sonar images.Compared with the result of global feature extraction,the result of this new method contains important texture information.The sonar image is recognized by sparse representation classification.The simulation result shows that the proposed method not only recognizes the sonar image in real time,but also improves the recognition accuracy.
出处
《中国航海》
CSCD
北大核心
2011年第3期1-4,共4页
Navigation of China
基金
国家自然科学基金(60904087)
关键词
船舶、舰船工程
声纳图像识别
灰度-梯度共生矩阵
稀疏表示
特征
纹理信息
ship
naval engineering
sonar image recognition
gray level-gradient co-occurrence matrix
sparse representation
feature
texture information
