现在的位置: 首页 > 综合 > 正文

复杂场景下的多目标跟踪 –心得

2018年02月23日 ⁄ 综合 ⁄ 共 1818字 ⁄ 字号 评论关闭

多目标跟踪一直都是cv领域一个让人蛋疼的问题。摸索了这么久,有些心得,可能有所欠缺,请多指教

有两种方案  

你可以用背景建模,先分出前景运动目标,但是这样分出来的目标会存在粘连,对于多人勾肩搭背的走或者前后的车辆行驶来说,简直是噩梦,怎么办哪?有方法!! 对于比较轻度的粘连,可以用用极度腐蚀,膨胀,这种简单的预处理的方法,还有就是药结合摄像头,在3维的空间来分割粘连,你想啊,现实中行驶的车辆是3维的对不对,但是我们得到的视频源确实把3维投影到2维的图像上,也就是损失了Z轴的信息,前后行驶的车辆就是这个情况,所以你就可以在3维中还原z轴的方向,这对于不同的车辆,z轴分量是 具有识别性的。

当然,你不喜欢这样做的话,也可以试试用检测器检测目标,为此你可能需要付出昂贵的时间代价,这可能需要你做优化,让你的牛逼的多线程和并行化处理颤抖起来吧!!小伙伴~

这里我真正觉得非常震撼多目标跟踪界的是这个叫ben的一 paper:

Stable Multi-Target Tracking in Real-Time Surveillance Video

先贴出他的abstract:

The majority of existing pedestrian trackers concentrate on maintaining the identities of targets, however systems for remote biometric analysis or activity recognition in surveil
lance video often require stable bounding-boxes around pedestrians rather than approximate locations. We present a multi-target tracking system that is designed specifically
for the provision of stable and accurate head location estimates. By performing data association over a slidingwindow of frames, we are able to correct many data as sociation errors and fill in gaps where observations are missed. The approach is multi-threaded
and combines asynchronous HOG detections with simultaneous KLT tracking and Markov-Chain Monte-Carlo Data Association (MCMCDA) to provide guaranteed real-time tracking in high definition video. Where previous approaches have used ad-hoc models for data association,
we use a more principled approach based on MDL which accurately models the affinity between observations. We demonstrate by qualitative and quantitative evaluation that the system is capable of providing precise location estimates for large crowds of pedestrians
in real-time. To facilitate future performance com
parisons, we will make a new dataset with hand annotated ground truth head locations publicly available.

这个总结起来说的话是hog adaboost 人头检测、klt运动特征点跟踪、和MCMCDA数据关联(原理诸如粒子滤波)。

这几个东西整起来是异常耗时的,所以需要你多线程 并行处理,值得注意的是人头检测会有一个人头面的估计,在人头面上检测人头,不仅可以 提高效率也可以提高准确度。

给出这个作者的homepage :http://ben.benfold.com/videos.html

看看他的人头检测

这么牛逼的算法让人想到美剧 person interest!!




抱歉!评论已关闭.