模式识别在激光雷达图像分析识别中的应用

• 作 者: 王小芬 张海娜 马玉芳
• 机 构: 商丘工学院,
• 出 处: 激光杂志
• 发表日期: 2017年02期
• 关键词: 模式识别 识别准确度 图像边缘 图像识别 图像分类 图像分析 支持向量机 特征分类 分类器 向量集

针对当前激光雷达图像识别精度低的难题,提出了基于向量机的激光雷达图像识别方法。首先对激光雷达图像进行预处理,重构模糊视觉特征,提取显著像素点特征,然后以激光雷达图像的显著像素点特征作为分类器输入数据,利用模式识别算法—支持向量机分类器进行模式识别,实现对激光雷达图像的特征分类和准确识别,最后实验结果表明,该方法的激光雷达图像分类识别准确度高于传统方法。

In view of the problem of low accuracy of current laser radar image recognition, a method of laser radar image recognition based on vector machine is proposed. Firstly, the preprocessing of the lidar image is carried out to reconstruct the fuzzy visual features and extract the features of significant pixels. Then, the significant pixel features of the lidar image are used as the input data of the classifier, and the pattern recognition algorithm Finally, the experimental results show that the accuracy of LIDAR image classification and recognition is higher than that of the traditional method.