| 利用行人运动特性的TLD快速检测算法 |
| |
| 引用本文： | 杨大伟,郭 超,吕伊鹏,程 凡.利用行人运动特性的TLD快速检测算法[J].大连民族学院学报,2017,19(1):36-39. |
| |
| 作者姓名： | 杨大伟 郭 超 吕伊鹏 程 凡 |
| |
| 作者单位： | 大连民族大学 机电工程学院,辽宁 大连 116605 |
| |
| 基金项目： | 大连金州新区科技创新主体支持计划(KJCX-ZTPY-2014-0005);中央高校基本科研业务费专项资金自主项目(DC201501089)。 |
| |
| 摘 要： | 针对TLD目标检测时需全局穷举搜索耗时较高的问题,提出一种基于行人运动特性的区域优化算法。该算法可有效预测行人在图像中的可能区域,从而减小检测计算复杂度,提高算法效率。经仿真分析表明,该算法在原有目标检测性能不变的情况下,检测耗时较原算法平均降低81.54 %,跟踪速率平均提升4倍,跟踪实时性明显提高。
|
| 关 键 词： | 行人检测 运动特性 TLD算法 区域优化 |
An Improved TLD Algorithm Based on Motion Characteristics of Pedestrian |
| |
| YANG Da-wei,GUO Chao,LV Yi-peng,CHENG Fan.An Improved TLD Algorithm Based on Motion Characteristics of Pedestrian[J].Journal of Dalian Nationalities University,2017,19(1):36-39. |
| |
| Authors: | YANG Da-wei GUO Chao LV Yi-peng CHENG Fan |
| |
| Affiliation: | School of Electromechanical Engineering, Dalian Minzu University, Dalian Liaoning 116605, China |
| |
| Abstract: | For the huge time-waste problem in original TLD algorithm with full-search object-detection, this paper proposes a regional optimization algorithm which is based on the motion characteristics of pedestrian. It can effectively predict the pedestrian’s position in the image so as to reduce object-detection computational complexity and improve efficiency of original TLD algorithm. Simulation analysis shows that the proposed algorithm can reduce 81.54 % detection time-cost in average and increase 4 times tracking-speed in average on the basis of the original TLD algorithm object-detection process, which obviously improved the real-time performance of object tracking compared with the original TLD algorithm. |
| |
| Keywords: | pedestrian detection motion characteristics of pedestrian TLD algorithm regional optimization |
|
| 点击此处可从《大连民族学院学报》浏览原始摘要信息 |
|
点击此处可从《大连民族学院学报》下载全文 |
|
