共查询到18条相似文献,搜索用时 78 毫秒
1.
Guo等人利用n个水平集方程构造n个区域提出一种改进的CV模型(简称MCV模型),该模型需要的迭代次数很少,提高了图像分割的效率,但其分割结果受初始曲线位置的影响较大,极易陷入局部最优,无法分割复杂图像,且利用传统的Heviside函数无法得到准确的均值信息,因此无法保证数值的稳定性。本文对MCV模型进行改进,先对图像进行预分割得到初始曲线以提高分割效率且能保证分割结果全局最优,构造新的符号函数取代传统的Heviside函数改进MCV模型以保证数值稳定性。对MR图像进行的分割实验表明,其在保证迭代次数较少的同时分割更加准确。 相似文献
2.
虹膜分割是虹膜识别系统中最重要的环节,其分割的好坏将影响虹膜识别的准确率,而虹膜识别也是最可靠的人体生物终身身份标志之一。因此,提出了基于水平集算法的虹膜分割算法。此算法是利用水平集隐式特点与圆形形状方程显式的特点相融合确保了演化曲线在演化过程中仍保持圆形,利用其思想分割内边缘。引入自适应面积项到形状约束的CV模型中用来约束外边缘。实验结果表明,尽管眼睛睁开有限、眼镜和睫毛及眼睑等遮挡以及成像设备形成图像的角度等问题,此模型仍能取得很好的分割效果。选用区域相互重叠度——DICE作为分割算法的评价指标,由实验数据可知,提出的算法对虹膜分割是有效的。 相似文献
3.
针对桥梁蜂窝麻面图像经常存在光照不均、多背景并存的干扰问题,提出了基于HSI颜色空间与灰度波动相结合的复杂桥梁蜂窝麻面的图像分割算法。首先,绘制S分量灰度变化曲线;其次,搜索曲线所有潜在的波峰波谷,并求相邻波峰波谷的高度差;然后,基于灰度像素个数差分值的标准差筛选出部分高度差;最后,基于部分高度差的标准差搜索最佳阈值完成图像的阈值分割。实验结果表明,与二维OTSU法、Niblack法、二维Tsallis熵法等几种算法相比,该算法的分割效果和实时性更好。 相似文献
4.
基于HSI颜色空间的X射线彩色图像分割 总被引:7,自引:0,他引:7
给出一种HSI颜色空间上基于HSI分量统计概率分布的彩色图像分割方法,并应用于X射线安全检查仪下危险品图像的分割。该方法将RGB彩色图像转换到HSI颜色空间,并以S分量为主要依据对图像进行粗分割,获得旅客包裹图像,然后,利用H和1分量的联合阈值进行细节分割获得疑似危险品的图像。在实际应用中表明该方法快速简单,满足实时性要求,是一种高效的自动算法。 相似文献
5.
本文讨论了彩色图象在空闻转换中的分割方法。在RGB空间中,由于R、G、B分量之闻有很高的相关性,直接利用这些分量不能得到所需的效果,探讨了经过正交变换后的分割,实验结果表明分割效果较好。HSI彩色空间和YCbCr彩色空间都具有把亮度分量与其它分量分离的优点。在样本集协方差矩阵的特征向量构成的空间中进行分割效果更好.表明肤色点经转换后在这一肤色空间中具有良好的聚类特性,在YCbCr空间中分割效果较差。在三种分割方法中,如果图象中含有大面积的与肤色十分相似的区域,分割的结果将不尽如人意,这可与人脸的其它特征相结合,最终实现肤色的正确分割。 相似文献
6.
模糊C均值(FCM)被广泛应用于彩色图像分割中,但传统的模糊C均值由于没有考虑空间信息,因此对噪声特别敏感。针对此问题,提出了一种在HIS颜色空间结合像素邻域空间信息的模糊聚类新方法。实验结果表明,此方法对高噪声图像有较好的处理结果。 相似文献
7.
本文讨论了彩色图象在空间转换中的分割方法。在RGB空间中,由于R、G、B分量之间有很高的相关性,直接利用这些分量不能得到所需的效果,探讨了经过正交变换后的分割,实验结果表明分割效果较好。HSI彩色空间和YCbCr彩色空间都具有把亮度分量与其它分量分离的优点。在样本集协方差矩阵的特征向量构成的空间中进行分割效果更好,表明肤色点经转换后在这一肤色空间中具有良好的聚类特性,在YCbCr空间中分割效果较差。在三种分割方法中,如果图象中含有大面积的与肤色十分相似的区域,分割的结果将不尽如人意,这可与人脸的其它特征相结合,最终实现肤色的正确分割。 相似文献
8.
一种彩色图像快速分割方法 总被引:3,自引:0,他引:3
提出一种基于HSI和FCM的彩色图像快速分割算法CISHF.首先将彩色图像从RGB色彩空间转换到HSI空间,然后联合利用S(饱和度)分量和 I(亮度)分量进行粗分割,最后针对H(色调)分量进行模糊聚类.根据色调数据的特点,修正了样本数据到聚类中心的距离计算公式,给出统计有效样本权重的算法,对于有效色调值进行样本加权聚类,加快了聚类速度.实验表明,CISHF算法的运算性能大大高于标准FCM算法,获得了较好的彩色图像分割效果. 相似文献
9.
10.
基于Mumford-Shah模型的快速水平集图像分割方法 总被引:78,自引:4,他引:78
该文对Chan-Vese提出的基于Mumford-Shah模型的水平集分割图像的算法做了两方面的改进:首先改进了C-V方法的偏微分方程,使得C-V方法可以快速计算出全局最优分割;其次,采用源点映射扫描方法来快速计算符号距离函数,克服了常规水平集方法中构造符号距离函数计算量大的缺点,并结合该文所提出的基于快速步进法生成符号表的方法,进一步提高了计算稳定性.两方面的改进提高了计算的速度和分割效果,试验统计结果显示,对于512×512的大幅图像,一般只需要10次左右的迭代就可以得到最优的分割效果.对合成图像、生物医学图像的分割结果表明了本文方法的稳健、快速. 相似文献
11.
针对有雾图像对比度差、能见度低的情况,结合HSI颜色空间特点,提出一种单幅图像去雾算法。首先,将有雾图像从RGB颜色空间转换到HSI颜色空间;然后,依据HSI颜色空间中色度、饱和度和亮度各分量受雾影响程度的差异,建立相应的去雾模型;最后,通过分析图像饱和度,得到饱和度模型中权重的取值范围,再对亮度模型中权重进行估计,从而实现去雾效果。与其他几种算法的实验结果比较表明,所提算法运算效率提高1倍左右。同时该算法能有效增强图像清晰度,能很好地运用于单幅图像去雾。 相似文献
12.
提出了一种基于HSI空间的假彩色异类传感器图像融合算法。该算法从边缘、色彩这两个图像的基本特征着手,在灰度融合图像的基础上,基于HSI空间,用色彩体现特殊传感器的独有信息。选用laplacian金字塔融合算法及一致性融合规则得到边缘灰度融合图像,并将其送至HSI空间的通道;根据需要,将某特殊传感器获取的图像,送至H通道进行调制;最后将HSI空间的融合图像变换到RGB空间显示。仿真结果显示:本文算法得到的彩色融合图像不仅保留了与灰度融合图像相近的空间分辨率,而且突出了独有信息,是一幅信息量比较完备的融合图像。同时,对于HSI三通道的选择是固定的,具有很好的算法稳健性。 相似文献
13.
原生质体细胞的显微图像具有边界模糊、内部分布不均匀的特点,利用传统分割方法较难取得理想分割效果.针对原生质体细胞的圆形特点,在快速水平集分割算法中加入圆形先验知识,提出一种新的基于圆形约束的快速水平集模型.为解决多个原生质体细胞分割问题,首先对图像进行预分割,然后利用多个水平集表示的圆形约束快速模型进行再分割.对传统快速水平集进行改进得到一种基于直方图统计的快速水平集模型,利用该模型进行预分割可以取得较好的效果.对多个不粘连细胞和多个粘连细胞,分别采用八链码跟踪法和随机霍夫圆检测法对预分割后的目标区域进行分裂.实验结果表明,本文快速水平集算法可以有效地解决单个及多个原生质体细胞分割问题. 相似文献
14.
针对煮糖过程蔗糖结晶图像的特点,采用颜色空间转换的方法,将图像从RGB颜色空间转换到HSI颜色空间,利用HSI颜色空间各分量相对独立性以及结晶颗粒和糖浆溶液的色调差异通过改进的大津法对H分量进行阈值分割,再通过数学形态学,中值滤波,孔洞填充,去除噪声颗粒,去除图像边界的非完整颗粒等方法进行后续处理,最终将结晶颗粒从复杂的图像信息中分割出来。实验证明该方法分割效果好,切实可行。 相似文献
15.
为实现对图像更好的阈值分割,针对原始色差变换存在的同色异值、异色同值的问题,基于HSI彩色空间色差变换,提出正余弦交叉法(SCCM)取参考点。该变换方法是根据各点色调正余弦值计算总体正、余弦均值,再由此求得对应的色调值。方法能够根据不同图像自适应地计算色差变换的参考点,使得变换后的灰度图有较好的灰度分布,为阈值分割提供更好的分割对象,提高分割质量。 相似文献
16.
改进的HSI空间形态学有噪彩色图像边缘检测 总被引:1,自引:0,他引:1
针对在RGB空间中很难有效区分颜色相似性问题,选择了更加符合颜色视觉特性的HSI颜色空间进行图像处理,提出了一种改进的形态学有噪彩色图像边缘检测方法,将开闭的迭代运算和双结构元多尺度运算应用到传统形态学梯度算子中,然后计算图像H、S、I三个分量的边缘信息,根据H、S、I所占比重对三分量进行加权融合得到彩色图像边缘.实验结果表明,该方法所检测的边缘符合人眼视觉特性,在抗噪声方面的效果比传统方法及其他多种方法更佳,能够更完整地保留原彩色图像的轮廓,计算量相对较小,有很好的实用性和通用性. 相似文献
17.
This article presents a method for classifying color points for automotive applications in the Hue Saturation Intensity (HSI)
Space based on the distances between their projections onto the SI plane. Firstly the HSI Space is analyzed in detail. Secondly
the projection of image points from a typical automotive scene onto the SI plane is shown. The minimal classes relevant for
driver assistance applications are derived. The requirements for the classification of the points into those classes are obtained.
Several weighting functions are proposed and a fast form of an euclidean metric is investigated in detail. In order to improve
the sensitivity of the weighting function, dynamic coefficients are introduced. It is shown how to compute them automatically
in order to get optimal results for the classification. Finally some results of applying the metric to the sample images are
shown and the conclusions are drawn.
Calin Rotaru is a PhD candidate at the Department of Computer Science, University of Hamburg, Germany. His PhD work focuses on the topic color machine vision for driver assistance systems and is supported by Volkswagen AG, Group Research Electronics. He graduated (2002) with the topic “Stereo Camera Based Object Recognition” for Driver Assistance Systems from the Faculty of Automation and Computer Science of the Technical University of Cluj-Napoca, Romania. His research interests include color machine vision, smart vision systems, multisensorial data fusion and vision in driver assistance systems. Thorsten Graf received the diploma (M.Sc.) degree in computer science and the Ph.D. degree (his thesis was on “Flexible Object Recognition Based on Invariant Theory and Agent Technology”) from the University of Bielefeld, Bielefeld, Germany, in 1997 and 2000, respectively. In 1997 he became a Member of the “Task Oriented Communication” graduate program, University of Bielefeld, funded by the German research foundation DFG. In June 2001 he joined Volkswagen Group Research, Wolfsburg, Germany. Since then, he has worked on different projects in the area of driver assistance systems as a Researcher and Project Leader. He is the author or coauthor of more than 40 publications and owns several patents. His research interests include image processing and analysis dedicated to advanced comfort/safety automotive applications. Dr. Jianwei Zhang is full professor and director of the Institute of Technical Aspects of Multimodal Systems, Department of Computer Science, University of Hamburg, Germany. He is one of the Chair Professors “Human-Computer Interaction” of the Department of Computer Science of Tsinghua University. He received his Bachelor (1986) and Master degree (1989) from the Department of Computer Science of Tsinghua University, and his PhD (1994) from the Department of Computer Science, University of Karlsruhe, Germany. His research interests include multimodal information processing, robot learning, service robots, smart vision systems and Embodied Intelligence. In these areas he has published over 120 journal and conference papers, six book chapters and two research monographs. He leads numerous basic research and application projects, including the EU basic research programs and the Collaborative Research Centre supported by the German Research Council. Dr. Zhang has received multiple awards including the IEEE ROMAN Best Paper 2002. 相似文献
Jianwei ZhangEmail: |
Calin Rotaru is a PhD candidate at the Department of Computer Science, University of Hamburg, Germany. His PhD work focuses on the topic color machine vision for driver assistance systems and is supported by Volkswagen AG, Group Research Electronics. He graduated (2002) with the topic “Stereo Camera Based Object Recognition” for Driver Assistance Systems from the Faculty of Automation and Computer Science of the Technical University of Cluj-Napoca, Romania. His research interests include color machine vision, smart vision systems, multisensorial data fusion and vision in driver assistance systems. Thorsten Graf received the diploma (M.Sc.) degree in computer science and the Ph.D. degree (his thesis was on “Flexible Object Recognition Based on Invariant Theory and Agent Technology”) from the University of Bielefeld, Bielefeld, Germany, in 1997 and 2000, respectively. In 1997 he became a Member of the “Task Oriented Communication” graduate program, University of Bielefeld, funded by the German research foundation DFG. In June 2001 he joined Volkswagen Group Research, Wolfsburg, Germany. Since then, he has worked on different projects in the area of driver assistance systems as a Researcher and Project Leader. He is the author or coauthor of more than 40 publications and owns several patents. His research interests include image processing and analysis dedicated to advanced comfort/safety automotive applications. Dr. Jianwei Zhang is full professor and director of the Institute of Technical Aspects of Multimodal Systems, Department of Computer Science, University of Hamburg, Germany. He is one of the Chair Professors “Human-Computer Interaction” of the Department of Computer Science of Tsinghua University. He received his Bachelor (1986) and Master degree (1989) from the Department of Computer Science of Tsinghua University, and his PhD (1994) from the Department of Computer Science, University of Karlsruhe, Germany. His research interests include multimodal information processing, robot learning, service robots, smart vision systems and Embodied Intelligence. In these areas he has published over 120 journal and conference papers, six book chapters and two research monographs. He leads numerous basic research and application projects, including the EU basic research programs and the Collaborative Research Centre supported by the German Research Council. Dr. Zhang has received multiple awards including the IEEE ROMAN Best Paper 2002. 相似文献