基于Krawtchouk矩的运动目标分类

摘要

针对视频监控系统,为了实现对运动目标的有效分类,提高其分类准确率,提出一种基于Krawtchouk矩不变量特征的运动目标分类技术.提取目标图像的低阶Krawtchouk矩不变量来描述目标,确保目标在平移、缩放、旋转条件下的不变性,选用K-means均值聚类算法进行数据聚类,达到准确分类的效果.通过实验验证了该方法的有效性,与应用Hu矩不变量进行分类的结果进行比较,表明了该方法具有较好的分类效果.%A moving target classification method based on Krawtchouk moment is proposed to achieve effective classification of moving targets, which is used in video surveillance system. In order to achieve accurate classification results, the low-level Krawtchouk moment invariants of the target images are extracted to ensure translation, scale and rotation invariance> then cluster data using K-means clustering algorithm. At last, the experiments demonstrate the effectiveness of this method, the result shows this method has a better effectiveness than the method with Hu moment invariants.