相关序列小目标图像背景杂波度量方法仿真  

Simulation of Measurement Method of Background Clutter in Correlation Sequence Small Target Image

作　　者：龙波 LONG Bo(School of Computer Science,Southwest Petroleum University,Chengdu Sichuan 610500,China)

机构地区：[1]西南石油大学计算机科学学院,四川成都610500

出　　处：《计算机仿真》2020年第9期377-381,共5页Computer Simulation

摘　　要：为了使图像背景的修复更为完整、效率更快,提出了一种相关序列小目标图像背景杂波度量方法,通过对图像背景的架构相似性进行度量,得到SSIM图,利用Canny算子边缘检测和梯度估算将SSIM图转换成梯度直方图,并获取其特征向量,提取目标图像背景的特征,构建图像背景特征空间,提取图像背景特征空间的相似度特征,估算不同区域背景和目标的架构相似度,获得相似度曲面,在曲面内将相似度特征与图像背景特征进行对比,进而完成对相关序列小目标图像背景杂波的度量。仿真结果证明,相较于传统方法,所提方法具有度量效率快与度量精准度高的优点。In order to make the restoration of image background more complete,this article presented a method to measure the background clutter of small target image in correlation sequence.At first,SSIM image was obtained by measuring the similarity between image background structures.Then,SSIM image was transformed into the gradient histogram by Canny operator edge detection and gradient estimation,and its feature vector was obtained.Moreover,the feature of target image background was extracted and its image background feature space was constructed.Moreo⁃ver,the similarity feature of image background space was extracted to estimate the similarity between different region⁃al backgrounds and target architectures,so as to find the similarity surface.Finally,the similarity feature was com⁃pared with image background feature in surface,and then the measurement of background clutter of small target image in correlation sequence was completed.Simulation results show that the proposed method has the faster measurement efficiency and higher measurement accuracy.

关 键 词：图像背景 杂波度量 相似度 特征空间 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]