摘要
针对光学伪装效果评价中的目标轮廓变形问题,提出了一种基于轮廓形变度的二值统计矩算法。首先,对伪装目标的原始背景图像进行二值化处理,然后对背景图像均匀分割,提取目标的轮廓特征,以此构造一个轮廓特征向量的二值统计矩阵。最后,采用欧氏距离与余弦归一化的方法对二值统计矩进行计算,获得目标的轮廓形变度值。实验结果表明,提出的二值统计矩算法可以有效地提取目标的轮廓特征;轮廓形变度指标达到了0.905±0.004和0.77±0.80;与传统Hu矩算法相比,主成分因子分别提高了27%和7%,算法效率提高了69.8%。该方法能够从目标轮廓的角度有效地评估光学迷彩伪装效果。
Aiming at the problem of target contour deformation in the effect evaluation of optical camouflage,this paper proposes a binary statistical moment algorithm based on contour deformation degree.Firstly,the original background image of a camouflage target is binarized,followed that the background image is evenly segmented,and then the contour features of the target are extracted,so that a binary statistical matrix of the contour feature vector is constructed.Finally,both the Euclidean distance and Cosine normalization are adopted to calculate the binary statistical moment,so as to obtain the contour deformation degree of the target.The experimental results show that,the proposed algorithm can effectively extract the contour features of the target;the contour deformation index reaches 0.905±0.004 and 0.77±0.80;compared with the traditional Hu moment algorithm,the principal component factor is increased by 27%and 7%respectively,and the algorithm efficiency is improved by 69.8%.The optical camouflage effect can be effectively evaluated from the perspective of the target contour.
作者
喻钧
刘昊阳
张云辉
胡志毅
初苗
YU Jun;LIU Haoyang;ZHANG Yunhui;HU Zhiyi;CHU Miao(School of Computer Science and Engineering,Xi'an Technological University,Xi'an 710021,China;PLA 32182,Beijing 100042,China;School of Art and Media,Xi'an Technological University,Xi'an 710021,China)
出处
《光子学报》
EI
CAS
CSCD
北大核心
2021年第6期188-196,共9页
Acta Photonica Sinica
基金
国家自然科学基金(No.51905407)。
关键词
光学伪装效果
目标轮廓
轮廓形变度
二值统计矩
Optical camouflage effect
Target contour
Contour deformation degree
Binary statistical moment
