基于灰度变化率的低对比度CT图像分割研究
|
摘要
整幅工业CT图像具有部分区域对比度低,灰度范围狭窄,灰度变化不明显等特点,针对传统的缺陷分割方法无法对低对比度区域的缺陷进行精确分割的问题,提出一种基于灰度变化率的缺陷分割方法.通过提取图像中某点的灰度值并计算该点与其周围邻域内点的平均灰度值的变化率、差值以及方差,选取符合图像分割条件范围内的点作为边界点,从而提取工业CT图像中低对比度区域缺陷的边界并进行分割.仿真实验表明,本文方法分割CT图像的缺陷准确率可达到95%,能够快速确定缺陷区域,并准确、有效地分割提取缺陷.
According to the characteristics of low contrast,narrow gray range and inconspicuous change of gray value in the part of industrial CT image,and the problem of difficult defect segmentation of low contrast region of the traditional segmentation method,a defect segmentation method is proposed based on the change rate of gray level.In order to determine and split the edge of defects in low contrast,the change rate of average gray value,difference and variance between the point and the neighborhood point are calculated,respectively.The points which accord with the condition of image segmentation are selected as boundary points,and the defect edge of low contrast area is identified and segmented in industrial CT image.Simulation results show that the proposed algorithm can increase the accuracy rate of segmentation to 95%,can quickly determine the defect area,and accurately and effectively segment and extract the defects.
According to the characteristics of low contrast,narrow gray range and inconspicuous change of gray value in the part of industrial CT image,and the problem of difficult defect segmentation of low contrast region of the traditional segmentation method,a defect segmentation method is proposed based on the change rate of gray level.In order to determine and split the edge of defects in low contrast,the change rate of average gray value,difference and variance between the point and the neighborhood point are calculated,respectively.The points which accord with the condition of image segmentation are selected as boundary points,and the defect edge of low contrast area is identified and segmented in industrial CT image.Simulation results show that the proposed algorithm can increase the accuracy rate of segmentation to 95%,can quickly determine the defect area,and accurately and effectively segment and extract the defects.
引文
[1]王秉欣,丛鹏.高精度工业CT三维图像的标的挖掘与分析[J].原子能科学技术,2013,47(1):137-140.
[2]卢艳平,王珏,喻洪麟.工业CT三维图像处理与分析系统[J].仪器仪表学报,2009,30(2):444-448.
[3]许新征,丁世飞,史忠植,等.图像分割的新理论和新方法[J].电子学报,2010,38(S1):76-82.
[4]TU S J,SHAW C C,CHEN L Y.Noise simulation in cone beam CT imaging with parallel computing[J].Physics in Medicine and Biology,2006,51(5):1283-1297.
[5]向新程,丛鹏,陈景运.二维直方图阈值法分割工业CT图像的研究[J].原子能科学技术,2007,41(3):347-350.
[6]何俊,葛红,王玉峰.图像分割算法研究综述[J].计算机工程与科学,2009,31(12):58-61.
[7]CHENG H D,CHEN Y H,JIANG X H.Thresholding using two-dimensional histogram and fuzzy entropy principle[J].Image Processing,IEEE Transactions,2000,9(4):732-735.
[8]MASOOLEH M G,ALI S,MOOSAVI S.An improved fuzzy algorithm for image segmentation[J].Proceedings of World Academy of Science,Engineering and Technology,2008,28(4),400-404.
[9]师文,朱学芳,朱光.基于形态学的MRI图像自适应边缘检测算法[J].仪器仪表学报,2013,34(2):408-414.
[10]聂守平,王鸣,刘峰.低对比度图像分割算法研究[J].中国激光,2004,31(1):89-91.
[11]陆剑锋,林海,潘志庚.自适应区域生长算法在医学图像分割中的应用[J].计算机辅助设计与图形学学报,2005,17(10):2169-2173.
[12]SHAFARENKO L,PETROU M,KITTLER J.Automatic watershed segmentationof randomly textured color images[J].IEEE Trans.Image Processing,1997,6(11):1530-1543.
[13]唐子淑,刘杰,别术林.基于CV模型的CT图像分割研究[J].CT理论与应用研究,2014,23(2):193-202.
[14]叶建平,郭李云,田毅.一种低对比度CT图像的血管分割方法[J].计算机系统应用,2015,24(2):184-188.
[15]刘金清,陈锟.一种CT医学图像分割新算法[J].微计算机应用,2011,32(10):7-12.
[16]DENG J W,TSUI H T.A fast level set method for segmentation of low contrast noisy biomedical images[J].Pattern Recognition Letters,2002,23(1):161-169.
[17]王栋,朱明.低对比度图像中改进的二维熵阈值分割法[J].仪器仪表学报,2004,25(S4):355-362.
[18]黄魁东,张定华,金炎芳.低对比度锥束CT图像缺陷分割算法[J].核电子学与探测技术,2009,29(1):132-136.
[19]张建军.基于区域灰度变化率的图像边界检测[D].昆明:云南师范大学,2010.
[20]吕哲,王福利,常玉清.一种改进的Canny边缘检测算法[J].东北大学学报(自然科学版),2007,28(12):1681-1684.
[21]王珏,陈教泽,谭辉,等.工业CT探测系统噪声特性研究[J].核电子学与探测技术,2010,30(7):929-934.
[2]卢艳平,王珏,喻洪麟.工业CT三维图像处理与分析系统[J].仪器仪表学报,2009,30(2):444-448.
[3]许新征,丁世飞,史忠植,等.图像分割的新理论和新方法[J].电子学报,2010,38(S1):76-82.
[4]TU S J,SHAW C C,CHEN L Y.Noise simulation in cone beam CT imaging with parallel computing[J].Physics in Medicine and Biology,2006,51(5):1283-1297.
[5]向新程,丛鹏,陈景运.二维直方图阈值法分割工业CT图像的研究[J].原子能科学技术,2007,41(3):347-350.
[6]何俊,葛红,王玉峰.图像分割算法研究综述[J].计算机工程与科学,2009,31(12):58-61.
[7]CHENG H D,CHEN Y H,JIANG X H.Thresholding using two-dimensional histogram and fuzzy entropy principle[J].Image Processing,IEEE Transactions,2000,9(4):732-735.
[8]MASOOLEH M G,ALI S,MOOSAVI S.An improved fuzzy algorithm for image segmentation[J].Proceedings of World Academy of Science,Engineering and Technology,2008,28(4),400-404.
[9]师文,朱学芳,朱光.基于形态学的MRI图像自适应边缘检测算法[J].仪器仪表学报,2013,34(2):408-414.
[10]聂守平,王鸣,刘峰.低对比度图像分割算法研究[J].中国激光,2004,31(1):89-91.
[11]陆剑锋,林海,潘志庚.自适应区域生长算法在医学图像分割中的应用[J].计算机辅助设计与图形学学报,2005,17(10):2169-2173.
[12]SHAFARENKO L,PETROU M,KITTLER J.Automatic watershed segmentationof randomly textured color images[J].IEEE Trans.Image Processing,1997,6(11):1530-1543.
[13]唐子淑,刘杰,别术林.基于CV模型的CT图像分割研究[J].CT理论与应用研究,2014,23(2):193-202.
[14]叶建平,郭李云,田毅.一种低对比度CT图像的血管分割方法[J].计算机系统应用,2015,24(2):184-188.
[15]刘金清,陈锟.一种CT医学图像分割新算法[J].微计算机应用,2011,32(10):7-12.
[16]DENG J W,TSUI H T.A fast level set method for segmentation of low contrast noisy biomedical images[J].Pattern Recognition Letters,2002,23(1):161-169.
[17]王栋,朱明.低对比度图像中改进的二维熵阈值分割法[J].仪器仪表学报,2004,25(S4):355-362.
[18]黄魁东,张定华,金炎芳.低对比度锥束CT图像缺陷分割算法[J].核电子学与探测技术,2009,29(1):132-136.
[19]张建军.基于区域灰度变化率的图像边界检测[D].昆明:云南师范大学,2010.
[20]吕哲,王福利,常玉清.一种改进的Canny边缘检测算法[J].东北大学学报(自然科学版),2007,28(12):1681-1684.
[21]王珏,陈教泽,谭辉,等.工业CT探测系统噪声特性研究[J].核电子学与探测技术,2010,30(7):929-934.