首页|基于再模糊和奇异值分解的局部模糊区域检测

基于再模糊和奇异值分解的局部模糊区域检测

扫码查看
针对局部模糊区域检测与分割,提出一种基于再模糊和奇异值分解的局部模糊检测方法.首先,对待检图像进行高斯模糊和图像连续分块,对分块图像进行奇异值分解;其次,对奇异值向量进行指数函数拟合,得到待测和再模糊图像的奇异值曲线,分别计算拟合曲线的积分并作差,将差值差异性作为描述模糊的特征值;最后,利用Canny边缘检测算法完成图像分割,并利用最大、最小滤波器和相关形态学处理对边缘内部区域进行填充.实验结果表明:该方法可以有效地完成局部模糊区域的检测和分割,同时对纹理平坦区域的识别同样有效.
Blur detection based on Re-blurred theory and singular value decomposition
To solve the problem of the detection and segment of local blur image,a local blur detection method based on re-blur theory and singular value decomposition is proposed.Firstly,the test image is pro-cessed by Gaussian Blur and continuously segmented,and the test image blocks are decomposed by singular value decomposition.Secondly,the singular value vector of every image block is fitted by exponential func-tion to get the singular value curve of the images to be measured and re-blured images.Then,the integral of singular value vector of image blocks are counted respectively and being subtracted,and the difference of subtraction of integral value is used as blur feature.Finally,the canny edge detection algorithm is used to complete the image segmentation and use the maximum and minimum filters and related morphological pro-cessing to deal with the internal area of the edge.The experiment results show that the method can effective-ly complete the detection and segmentation of local blur image,and it is also effective in the identification of flat regions of texture.

blur detectionre-blur theorysingular value decompositioncurve fitting integralblur feature

王奎奎、彭国晋、侯义锋、吕美妮、杨秋慧

展开 >

梧州学院,广西机器视觉与智能控制重点实验室,广西 梧州 543002

模糊检测 再模糊理论 奇异值分解 拟合曲线积分 模糊特征

国家自然科学基金广西自然科学基金广西自然科学基金广西壮族自治区创新驱动发展专项广西壮族自治区高等学校中青年教师科研基础能力提升项目广西壮族自治区高等学校中青年教师科研基础能力提升项目

620022682018GXNSFBA2811732021JJB170060桂科AA118180862019KY06922019-KY0686

2024

信息技术
黑龙江省信息技术学会 中国电子信息产业发展研究院 中国信息产业部电子信息中心

信息技术

CSTPCD
影响因子:0.413
ISSN:1009-2552
年,卷(期):2024.(3)
  • 10