«上一篇/Previous Article|本期目录/Table of Contents|下一篇/Next Article»

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

28

期数:

2018年03期

页码:

1-5

栏目:

智能、算法、系统工程

出版日期:

2018-03-10

文章信息/Info

Title:

A Saliency Fusion Detection Method Based on Image Features from Different Sensors

文章编号:

1673-629X(2018)03-0001-05

作者:

李海波1 ; 曹云峰2 ; 丁萌3 ; 庄丽葵2

1. 南京航空航天大学 自动化学院,江苏 南京 210016;
2. 南京航空航天大学 航天学院,江苏 南京 210016;
3. 南京航空航天大学 民航学院,江苏 南京 210016

Author(s):

LI Hai-bo1 ; CAO Yun-feng2 ; DING Meng3 ; ZHUANG Li-kui2

1. School of Automation Engineering,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China;
2. School of Astronautics,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China;
3. School of Civil Aviation,Nanjing University of Aeronautics and Astronautics,Nanjing 210016,China

关键词:

异源图像; 特征提取; 图像分割; 显著性融合; 目标检测

Keywords:

sensor images; feature extraction; image segmentation; saliency fusion; target detection

分类号:

TP301.6

DOI:

10.3969/ j. issn.1673-629X.2018.03.001

文献标志码:

A

摘要:

基于光学图像特征的显著性检测在光照条件发生变化的情况下易受到影响,为保证在深空探测等特殊场景中进行准确检测,研究了一种基于光学图像与深度图的显著性融合检测方法。 对于经过配准的光学图像与深度图,分别计算光学图像的显著性图与深度图的显著性图,将计算得到的两类显著性图进行融合,确定所包含的共同显著性区域,消除单类显著图计算过程中存在的异常显著区域,形成新的显著性图。 根据形成的显著性图,按设定阈值转化为二值图,确定显著性区域。 在光学图像中利用显著性区域信息确定目标区域,完成目标所在区域检测。 实验结果表明,该方法可以有效检测出显著性目标所在区域,与其他常用方法相比具有较高的检测准确率。

Abstract:

Saliency detection based on optical image characteristics is easily affected under the condition of light changes. To detect accurately in special scenario of deep space,we study a saliency fusion detection method based on optical image and depth map. For the registered optical image and depth map,the saliency maps of optical image and depth map are calculated respectively and then merged to determine the common salient areas and eliminate the abnormal salient areas in the calculation on single image. A new saliency map is generated and converted into binary image according to the setting threshold to identify significant areas. The significant region information is used to determine target areas on optical image. So far,the target region detection is finished. The experiments show that this method can effectively detect the saliency target areas with high detection accuracy compared with other common methods.

常用功能

更新日期/Last Update: 2018-04-03