Nuclear support vector machine target classification method based on millimeter wave radar point cloud features

The invention discloses a nuclear support vector machine target classification method based on millimeter-wave radar point cloud features, and the method comprises the steps: firstly carrying out thepreprocessing of original radar point cloud data, and removing point clouds outside a radar detection region; and clustering the target point clouds into one class through a clustering algorithm so asto eliminate noise point clouds; then constructing a feature vector composed of 11 features in combination with the point cloud features of the target, and performing training and testing by adoptinga kernel support vector machine classifier, thereby realizing target classification. Compared with a traditional target classification method based on millimeter-wave radar, the method has higher recognition accuracy, and has important practical significance for researching the perception ability of automatic driving. 本发明公开了一种基于毫米波雷达点云特征的核支持向量机目标分类方法，该方法首先对原始的雷达点云数据进行预处理，将雷达探测区域外的点云剔除。接着通过聚类算法将目标点云聚为一类，从而剔除噪声点云。然后结合目标的点云特点