• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>International Conference on Applications and Techniques in Cyber Intelligence >Target Recognition Based on Kinect Combined RGB Image with Depth Image
【24h】

Target Recognition Based on Kinect Combined RGB Image with Depth Image

机译：基于Kinect组合具有深度图像的RGB图像的目标识别

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Target recognition is the most important premise in detection and tracking of human objects. Using the global or local features to identify the human body is the mainstream way to track human body. This paper proposes a system of target recognition based on Kinect which combined the global and local features. Firstly, the RGB image is obtained by the Kinect. And then the people were separated from the background in the RGB image. After that the RGB image is transformed into a histogram. Then, the depth image was obtained by the Kinect. The spatial coordinates of bone nodes were obtained by the Kinect bone tracking technology such as the head, shoulder center, spine, and hip center of the human body. The Euclidean distance of the bone points in space were calculated. Finally, the results of each step were combined as a basis for recognition. Our method can reduce the influence of external light, light and other conditions on recognition.
机译：目标识别是人类物体检测和跟踪中最重要的前提。使用全局或本地特征来识别人体是跟踪人体的主流方式。本文提出了一种基于Kinect的目标识别系统，其组合全局和局部特征。首先，RGB图像通过Kinect获得。然后人们与RGB图像中的背景分开。之后，RGB图像被转换成直方图。然后，通过Kinect获得深度图像。骨节点的空间坐标是通过诸如人体的头部，肩部，脊柱和臀部中心的Kinect骨跟踪技术获得。计算空间中骨点的欧几里德距离。最后，将每个步骤的结果合并为识别的基础。我们的方法可以减少外部光，光线和其他条件对识别的影响。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号