针对纸张目标的带结构约束压缩跟踪算法

发布时间：2018-05-04 05:11

本文选题：目标跟踪 + 压缩感知　；参考：《计算机辅助设计与图形学学报》2017年06期

【摘要】：针对传统的视觉跟踪算法对于缺少显著纹理细节的空白纸张很难进行鲁棒跟踪,特别是目标出现大的旋转时跟踪效果不理想的问题,提出一种改进的带有结构约束的压缩跟踪算法.该算法针对样本集利用直线检测和直线夹角测定对样本集进行约束判断,排除了特征值近似而明显错误的样本块,减少了样本搜索空间;同时使得能够跟踪具有大幅度旋转的目标样本,还得到纸张顶角的准确位置,从而为纸张的整体定位奠定良好基础.实验结果表明,文中算法是有效的,能够很好地跟踪这类存在大幅度旋转的、无纹理特征的空白纸张目标.
[Abstract]:In view of the traditional visual tracking algorithm, it is very difficult for the blank paper without significant texture details to be robust tracking, especially for the problem that the tracking effect is not ideal when the target has a large rotation. An improved compression tracking algorithm with structural constraints is proposed. The algorithm uses straight line detection and linear angle measurement to judge the sample set, eliminates the sample block which is similar to the eigenvalue and obviously wrong, and reduces the sample search space. At the same time, it can track the target samples with large rotation and get the accurate position of the top angle of the paper, thus laying a good foundation for the overall positioning of the paper. Experimental results show that the proposed algorithm is effective and can be used to track this kind of blank paper targets with large rotation and no texture features.
【作者单位】：长安大学信息工程学院;
【基金】：中央高校基本科研业务费专项资金(310824173401)
【分类号】：TP391.41

【相似文献】