激光雷达图像分类的模式识别研究

Research on pattern recognition of laser radar image classification

激光雷达图像的类型众多,而且类型之间的相似度大比较大,当前激光雷达图像分类方法存在分类时间、分类误差大等不足,为了改善激光雷达图像分类效果,提出了基于模式识别技术的激光雷达图像分类方法。首先提取基于激光雷达图像的特征向量,然后采用模式识别技术进行激光雷达图像的分类,最后与当前它激光雷达图像分类方法进行了对比测试。本文方法的激光雷达图像分类正确率高于95%,远远高于其它方法的激光雷达图像分类正确率,改善了激光雷达图像分类效率,具有更加广泛的应用前景。

There are many types of laser radar images,and the similarity between types is large,and the current classification methods of lidar image have some shortcomings,such as long classification time and the large classification errors. In order to improve the classification effect of laser radar images,a method of classification of laser radar images based on pattern recognition technology is proposed. First,the feature vectors based on laser radar images are extracted,and then the pattern recognition technology is used to classify the lidar images. Finally,the comparison test is made with the current lidar image classification method. The classification accuracy of the laser radar image in this paper is higher than 95%,which is far higher than that of other methods. The classification efficiency of laser radar image is improved,and it has a wider application prospect.