参考:
1)时间同步器
https://blog.csdn.net/u012700322/article/details/51784412
https://blog.csdn.net/fana8010/article/details/23880067(这篇文章是上一篇的原文)
2)基于策略的同步器
https://blog.csdn.net/muyiyushan/article/details/118418040(官网翻译,第7部分)
http://wiki.ros.org/message_filters#Policy-Based_Synchronizer_.5BROS_1.1.2B-.5D(官网,第7部分)
以下是针对参考1)中的第一篇内容:
两句等价成一句,但为什么要分开成两句呢?因为只使用一句只能读取一个输入传给回调函数,所以需要使用两句,先用一句分开将多个数据封装好,然后再用一句将多个数据传给回调函数。
boost bind
The TimeSynchronizer filter synchronizes incoming channels by the timestamps contained in their headers
PointCloud2ROS点云数据消息结构:header里面有个stamp,比较不同数据源的stamp来进行时间同步。
以下是针对参考2)中的解释:
typedef message_filters::sync_policies::ApproximateTime<sensor_msgs::PointCloud2, sensor_msgs::PointCloud2> SyncPolicyT;
//定义策略,message_filters::sync_policies::ApproximateTime策略使用一个基于时间戳的自适应算法匹配消息,匹配两个点云
void DualLidarCalib::start()
{
// 订阅器 sub and pub
parent_cloud_sub_ = new message_filters::Subscriber<sensor_msgs::PointCloud2>(nh_, parent_cloud_topic_, 10);
child_cloud_sub_ = new message_filters::Subscriber<sensor_msgs::PointCloud2>(nh_, child_cloud_topic_, 10);
//点云同步器 ApproximateTime takes a queue size as its constructor argument, hence MySyncPolicy(100)
cloud_synchronizer_ = new message_filters::Synchronizer<SyncPolicyT>(SyncPolicyT(100), *parent_cloud_sub_, *child_cloud_sub_);
//调用回调函数
cloud_synchronizer_->registerCallback(boost::bind(&DualLidarCalib::cloudCallback, this, _1, _2));
//registerCallback()输入为const boost::shared_ptr<M const>&
}
void DualLidarCalib::cloudCallback(const sensor_msgs::PointCloud2::ConstPtr& in_parent_cloud_msg,
const sensor_msgs::PointCloud2::ConstPtr& in_child_cloud_msg)
{
...
}
boost::bind()理解:
1)https://blog.csdn.net/holyjw/article/details/89883363(只用看“1. 通过functions和function pointers使用bind”)
2)https://blog.csdn.net/bj318318/article/details/116200356(只用看“1介绍”)