摘要
为了获得理想的高能射线成像图像变化检测结果,提出了高斯混合模型的高能射线成像图像变化自动检测方法。首先采用mean shift算法对高能射线成像图像进行平滑处理,去除噪声,留存图像细节信息,并采用对数比值法分析图像中变化区域和未变化区域的概率统计分布,然后采用过四个高斯混合模型拟合差异图像的概率统计分布情况,并利用判断准则判断像素的归属,从而实现高能射线成像图像变化检测。结果表明:所提方法具备较好的图像平滑处理效果和图像变化检测效果,且检测性能较好,适用范围较大,对不同复杂度图像均具备较高检测效率。
In order to get ideal image change detection results of high-energy ray,an automatic change detection method of high-energy ray image based on Gaussian mixture model is proposed.Starting with the smoothing of highenergy ray image through mean shift algorithm,the noise is removed and the image details are remained.The logarithmic ratio method is used to analyze the probability and the statistical distribution of changed and unchanged areas in image.Then,the four Gaussian mixture models are adopted to fit the probability and the statistical distribution of difference image,and the judgment criteria is utilized to determine the pixels attribution,so as to achieve high-energy ray image change detection.The results show that the proposed method has good image smoothing effect and image change detection effect,the detection performance is good,the scope of application is large,and it has high detection efficiency for different complexity images.
作者
赵治月
崔茂齐
ZHAO Zhiyue;CUI Maoqi(Department of Mechanical and Electrical Engineering,Cangzhou Normal University,Cangzhou Hebei 061001,China)
出处
《激光杂志》
北大核心
2020年第8期115-118,共4页
Laser Journal
基金
河北省研究生创新资助项目(No.CXZZBS2019028)
沧州市自然科学(No.187000003)
沧州市重点研发计划指导项目(No.183102004)。
