首页> 外文会议>Third annual symposium on logic in computer science >Using the condensation algorithm to implement tracking for mobilerobots
【24h】

Using the condensation algorithm to implement tracking for mobilerobots

机译:使用凝聚算法实现对移动机器人的跟踪

获取原文
获取原文并翻译 | 示例

摘要

The detection of objects in every frame of a sequence is often notnsufficient for scene interpretation. Tracking can increase thenrobustness, especially when occlusions occur or when objects temporallyndisappear. The standard approach for tracking is to use a Kalman filternfor every object. This, however requires the use of a high complexitynmanagement system to deal with the multiple hypotheses necessary tontrack all anticipated objects. We present a stochastic approach which isnbased on the CONDENSATION algorithm-conditional density propagation overntime-that is capable of tracking multiple objects with multiplenhypotheses in range images. A probability density function describingnthe likely state of the objects is propagated over time using a dynamicnmodel. The measurements influence the probability function and allow thenincorporation of new objects into the tracking scheme. Additionally, thenrepresentation of the density function with a fixed number of samplesnensures a constant running time per iteration step. Results on differentndata sources are shown for mobile robot applications
机译:序列的每个帧中的对象检测通常不足以进行场景解释。跟踪可以增强鲁棒性,尤其是在发生遮挡或物体暂时消失时。跟踪的标准方法是对每个对象使用卡尔曼滤波器。但是,这需要使用高复杂度的管理系统来处理多种假设,从而必须跟踪所有预期对象。我们提出了一种基于CONDENSATION算法的随机方法-有条件的密度随时间的传播-能够跟踪距离图像中具有多个假设的多个对象。使用动态模型随时间传播描述对象的可能状态的概率密度函数。测量结果影响概率函数,然后允许将新对象合并到跟踪方案中。此外,用固定数量的样本表示密度函数可确保每个迭代步骤的运行时间恒定。显示了移动机器人应用程序在不同数据源上的结果

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