Cooperative moving-object tracking with multiple mobile sensor nodes — Size and posture estimation of moving objects using in-vehicle multilayer laser scanner
Cooperative moving-object tracking with multiple mobile sensor nodes — Size and posture estimation of moving objects using in-vehicle multilayer laser scanner
复制标题
多个移动传感器节点的协作运动物体跟踪——使用车载多层激光扫描仪估计运动物体的尺寸和姿态
DOI：
10.1109/icit.2016.7474726
复制
发表时间：
2016
期刊：
影响因子：
--
通讯作者：
Kazuhiko Takahashi
中科院分区：
文献类型：
--
作者：
Shota Kanaki;Koki Fujishita;M. Hashimoto;Ryohei Murabayashi;Kimiaki Inui;Kazuhiko Takahashi
我们提出了多个移动的传感器节点，每个配备了多层激光扫描仪的合作移动目标跟踪。在激光扫描图像中的运动目标检测的二值化的占有网格的方法。传感器节点将检测到的移动物体的信息发送到中央服务器，中央服务器使用贝叶斯滤波器估计这些物体的大小，位置和速度，并将估计值反馈给传感器节点。可靠的一对一（或一对多）匹配的跟踪对象和有关的移动对象的测量是通过基于规则的数据关联。通过两个传感器节点对车辆和人体的跟踪实验，验证了该方法的有效性。
We present cooperative moving object tracking with multiple mobile sensor nodes, each equipped with a multilayer laser scanner. Moving objects in the laser-scanned images are detected by a binarized occupancy grid method. The sensor nodes send the information of the detected moving objects to a central server, which estimates the size, position, and velocity of those objects using a Bayesian filter and feeds the estimates back to the sensor nodes. Reliable one-to-one (or one-to-many) matching of the tracked objects and the measurements related to the moving objects is achieved by rule-based data association. Experimental results of tracking car and people by two sensor nodes show the effectiveness of the proposed method.