基于数据挖掘的水下激光图像识别技术  

Underwater Laser Image Recognition Technology Based on Data Mining

作　　者：凤祥云[1,2] 卢辉斌[3] 

机构地区：[1]河北师范大学,石家庄050024 [2]河北建材职业技术学院,河北秦皇岛066000 [3]燕山大学,河北秦皇岛066004

出　　处：《激光杂志》2016年第1期55-58,共4页Laser Journal

基　　金：2014河北省信息化战略研究课题(X2014-032)

摘　　要：为了提高水下激光图像识别精度,提出一种基于数据挖掘的水下激光图像识别方法。首先收集水下激光图像,提取其Gabor特征,并采用主成分分析法对特征进行选择,消除冗余特征,然后根据特征对训练样本进行处理,并输入到相关向量机进行学习,建立水下激光图像识别分类器,最后采用具体水下激光图像进行仿真对比测试。测试结果表明,本文方法可以提高水下激光图像的识别率,而且获得较快的水下激光图像识别速度,具有较高的实际应用价值。In order to improve the accuracy of underwater laser image recognition, a new method of underwater laser image recognition based on data mining is proposed. Firstly, underwater laser images were collected, and the Gabor features were extracted, and the features were selected by principal component analysis to eliminate redundancy features, and then training samples were processed according to selected features, and then the training samples are input to the relevance vector machine to learn and establish recognition classifier of underwater laser image, and lastly, multi- ple underwater laser images are used to carried out simulation and comparison test. The results show that the proposed method can improve the recognition rate and obtain higher recognition speed of the underwater laser image, which has high practical application value.

关 键 词：水下激光图像 特征提取 主成分分析 图像分类器 相关向量机 

分 类 号：TN91[电子电信—通信与信息系统]