交比不变的Camshift跟踪方法

发布时间：2018-05-28 19:34

本文选题：目标跟踪 + Camshift跟踪　；参考：《光学精密工程》2016年04期

【摘要】：为了提高Camshift跟踪方法在复杂环境下的跟踪性能,应用被跟踪目标内部各特征像素间的交比不变原理,提出了一种改进的Camshift跟踪方法。该方法通过分析被跟踪的目标模型,计算出其内部各特征像素间的坐标关系;将内部数据间的交比不变量作为所提出的跟踪方法的约束条件,对跟踪错误的像素点进行校正,并将跟踪过程中连续两帧图像的内部特征像素间的距离比作为跟踪效果的评判标准。用改进的Camshift跟踪方法分别对标准测试视频内的视频信息和实际拍摄的视频信息进行了测试。结果显示,该方法在两种复杂环境实验条件下,跟踪目标的距离偏差都能保持在15pixel以内,对单帧图像平均处理时间在20ms以内。试验结果表明,该方法对复杂环境下的目标具有良好的跟踪效果,跟踪性能稳定,跟踪效率高,可以满足跟踪系统实时性的要求。
[Abstract]:In order to improve the tracking performance of the Camshift tracking method in complex environments, an improved Camshift tracking method is proposed by applying the principle of intersection invariance among the feature pixels within the target being tracked. By analyzing the target model, the coordinate relation of each feature pixel is calculated, and the intersecting invariant of the internal data is used as the constraint condition of the tracking method, and the tracking error pixel is corrected. The distance ratio between the internal feature pixels of two successive frames in the tracking process is taken as the criterion of tracking effect. The improved Camshift tracking method is used to test the video information in the standard test video and the actual video information taken. The results show that the range deviation of the tracking target can be kept within 15pixel and the average processing time for a single frame image is within 20ms under two kinds of complex environment experiments. The experimental results show that the method has a good tracking effect on the target in complex environment, stable tracking performance and high tracking efficiency. It can meet the real-time requirements of the tracking system.
【作者单位】：中北大学机械与动力工程学院;
【基金】：山西省科技攻关项目(No.20130321005-04)
【分类号】：TP391.41

【相似文献】