引用本文: | 王鲲鹏,张小虎,朱肇昆,等.基于形状熵差的相似多目标检测方法.[J].国防科技大学学报,2010,32(3):11-15.[点击复制] |
WANG Kunpeng,ZHANG Xiaohu,ZHU Zhaokun,et al.A New Method for Similar Multi-target Detection Based on Shape Entropy Difference[J].Journal of National University of Defense Technology,2010,32(3):11-15[点击复制] |
|
|
|
本文已被:浏览 7090次 下载 5924次 |
基于形状熵差的相似多目标检测方法 |
王鲲鹏, 张小虎, 朱肇昆, 于起峰 |
(国防科技大学 航天与材料工程学院,湖南 长沙 410073)
|
摘要: |
为实现靶场多目标图像的目标自动检测,利用多目标形状的相似性,提出一种基于形状熵差的目标检测新方法。提出了形状熵差的概念,将目标形状模型引入局部熵计算中,由局部内熵与局部外熵之差的极值点确定目标位置。基于此设计并实现了一种新的目标自动检测方法。通过实验证明了算法的有效性和鲁棒性。相比较于常见的一些目标检测算法,该方法具有更好的抗干扰能力和环境光照变化适应能力,可以应用于靶场多目标图像的目标检测,也可应用于其他类似的目标检测问题。 |
关键词: 靶场图像 光学测量 目标检测 多目标 熵 |
DOI: |
投稿日期:2009-12-11 |
基金项目:国家自然科学基金资助项目(171085) |
|
A New Method for Similar Multi-target Detection Based on Shape Entropy Difference |
WANG Kunpeng, ZHANG Xiaohu, ZHU Zhaokun, YU Qifeng |
(College of Aerospace and Materials Engineering, National Univ. of Defense Technology, Changsha 410073, China)
|
Abstract: |
In order to realize auto target detection of multi-target image of shooting ranges, the similarity of targets' shape was utilized, and a new method for auto similar target detection based on shape entropy difference was proposed. First, the concept of shape entropy difference was presented, which merged shape information into the calculation of entropy, and the target was detected by searching for the extreme point of entropy difference. Then, an auto target detection method based on shape entropy difference was realized. Finally, the validity and robustness of the algorithm was approved by experiment. Compared with the common algorithms of target detection, the method proposed here has better tolerance capability against noise and illumination variety. The method can be used for target detection of image of shooting ranges and other similar applications. |
Keywords: image of shooting ranges optical measurement target detection multi-target entropy |
|
|
|
|
|