融合时空上下文的复杂背景下多运动目标检测

计算机工程与科学 ›› 2020, Vol. 42 ›› Issue (12): 2186-2192.

融合时空上下文的复杂背景下多运动目标检测

张寅1,2，蔡旭阳2，许倩倩2，闫钧华1,2，苏恺2，张琨2

(1.南京航空航天大学空间光电探测与感知工业和信息化部重点实验室,江苏南京 211106;

2.南京航空航天大学航天学院，江苏南京 211106）

收稿日期:2019-10-10 修回日期:2020-05-22 接受日期:2020-12-25 出版日期:2020-12-25 发布日期:2021-01-05
基金资助:
国家自然科学基金（61705104）；江苏省自然科学基金(BK20170804)；国防科技创新特区项目支持

Multiple moving targets detection under complex background by integrating spatio-temporal context#br# #br# #br#

ZHANG Yin1,2,CAI Xu-yang2,XU Qian-qian2,YAN Jun-hua1,2,SU Kai2,ZHANG Kun2

(1.Key Laboratory of Space Photoelectric Detection and Perception (Nanjing University of Aeronautics and Astronautics),

Ministry of Industry and Information Technology,Nanjing 211106;

2.College of Astronautics,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China)

Received:2019-10-10 Revised:2020-05-22 Accepted:2020-12-25 Online:2020-12-25 Published:2021-01-05

摘要/Abstract

摘要： 针对目标进、出视场和被部分遮挡情况下检测率低的问题，提出一种联合时空上下文的多运动目标检测算法STC-MMTD。首先，利用时间上下文信息，基于前后向运动历史图提取候选目标区域；然后，利用空间上下文信息和目标表观信息，通过基于稀疏编码的CRF模型计算目标置信度图；最后，计算候选目标区域的目标置信度，检测出多运动目标。实验结果表明，所提算法具有良好的检测性能，在保证较高定位精度的同时，查全率、查准率和F测度均高于其他多目标检测算法的。

关键词: 多运动目标检测, 复杂背景, 时空上下文, 前后向运动历史图

Abstract: In order to solve the problem of low detection rate when targets enter/leave the field of view, or targets are partially occluded, a multi moving targets detection algorithm based on spatio- temporal context is proposed. Firstly, the time context information is used to extract the candidate target region, which is based on the forward and backward motion history map. Secondly, the spatial context information and the target apparent information are used to calculate the target confidence map, which is based on the sparsely encoded CRF model. Finally, the target confidence of the candidate target region is calculated to detect the multiple moving targets. The experimental results show that the proposed algorithm has good detection performance, and it has higher recall rate, precision and F measure than other multi-target detection algorithms under the condition of guaranteeing high positioning accuracy.

Key words: multiple moving targets detection, complex background, spatio-temporal context, forward and backward motion history map , ,

张寅, 蔡旭阳, 许倩倩, 闫钧华, 苏恺, 张琨 . 融合时空上下文的复杂背景下多运动目标检测[J]. 计算机工程与科学, 2020, 42(12): 2186-2192.

ZHANG Yin, CAI Xu-yang, XU Qian-qian, YAN Jun-hua, SU Kai, ZHANG Kun. Multiple moving targets detection under complex background by integrating spatio-temporal context#br# #br# #br#[J]. Computer Engineering & Science, 2020, 42(12): 2186-2192.

[1]	黄云明1，张晶1,2，喻小惠1，陶涛3，龚力波4. 改进STC和SURF特征联合优化的目标跟踪算法[J]. 计算机工程与科学, 2019, 41(10): 1795-1802.
[2]	张晶，王旭,范洪博. 自适应目标变化的时空上下文抗遮挡跟踪算法[J]. 计算机工程与科学, 2018, 40(09): 1653-1661.
[3]	苏相阁，张为. 强光干扰下的火焰前景提取技术研究[J]. 计算机工程与科学, 2017, 39(11): 2060-2065.
[4]	王培玉1,2,李峰1,2,周书仁1,2，廖卓凡1，2. 复杂码头环境下的船舶检测与跟踪算法[J]. 计算机工程与科学, 2017, 39(05): 992-998.
[5]	张武，黄帅，汪京京，刘连忠. 复杂背景下小麦叶部病害图像分割方法研究[J]. J4, 2015, 37(07): 1349-1354.
[6]	宋海声，刘平和，王全州，赵学深，王海燕. 基于人体骨骼和深度图像信息的指尖检测方法[J]. J4, 2014, 36(09): 1788-1794.
[7]	王志勇刘文奇. 基于数学形态学的快速运动物体检测[J]. J4, 2005, 27(1): 27-28.