Multi.scale fused edge detection algorithm based on conflict redistribution DSmT
QIAO Kui.xian1, YIN Shi.bai1，QU Sheng.jie2*
1.School of Computer Science and Engineering, Xian Technological University, Xi’an Shaanxi 710032, China;
2.School of Automation, Northwestern Polytechnical University, Xi’an Shaanxi 710072,China
Abstract:
Single-scale edge detection operator itself is sensitive to noise which leads to little difference between the real and false edge, so the edge detected by it is not accurate, because ground object character is complex and thin ground object is intermingled with noise in real environment. So a novel multi-scale fused edge detection algorithm based on conflict redistribution DSmT is proposed in this paper. First multi-scale edge measure is extracted and then evidence theory is brought in. The basic belief assignment of multi-scale edge measure is constructed by a new method of bidirectional exponent and then fused by conflict redistribution DSmT combination rule. At last edge points are extracted by multiple thresholds. Simulation with both optical and SAR images shows that the edge detection method of this paper suppresses noise effectively, meanwhile preserving rich details.Single.scale edge detection operator itself is sensitive to noise, which leads to little difference between the real and false edge, so the edge detected by it is not accurate, because ground object character is complex and thin ground object is intermingled with noise in real environment. Therefore, a new multi.scale fused edge detection algorithm based on conflict redistribution DSmT was proposed in this paper. First, multi.scale edge measure was extracted and then evidence theory was brought in. The basic belief assignment of multi.scale edge measure was constructed by a new method of bidirectional exponent and then fused by conflict redistribution DSmT combination rule. At last, edge points were extracted by multiple thresholds. The simulation with both optical and Synthetic Aperture Radar （SAR） images shows that the edge detection method of this paper suppresses noise effectively, while preserving rich details.Key words:
edge detection；Conflict Redistribution (CR)；multi.scale edge measure；Basic Belief Assignment (BBA)；evidence theory
0 引言
图像边缘检测是图像匹配、识别与分析等领域的基础课题，国内外在该领域展开大量的研究［1-6］。在复杂场景的实际应用中，图像往往存在负载地物并且容易受到噪声干扰，信噪比较低，单一尺度的Canny、Sobel和Log等算子提取的真实边缘与噪声点测度差异小，导致细小地物与高频噪声相互掺杂，边缘检测精度低。
边缘检测算子对噪声的鲁棒性和定位精度是相互矛盾的，小尺度算子有利于边缘定位，但对噪声极为敏感；大尺度算子虽然对噪声鲁棒性好，但边缘定位精度差，有时会丢失某些局部细节。多尺度边缘检测算法被证明是有效的［7-8］，在小尺度上可以对景象细节进行检测，在大尺度上能很好地抑制噪声，多个尺度融和后提取边缘，就可以在保留景象细节的基础上提高算法对噪声的鲁棒性。由于受到成像机理以及噪声的影响，任何边缘检测算子都存在不确定性，多尺度边缘测度之间也存在一定的冲突，证据推理理论被证明可以更好地处理具有不确定、冲突和模糊的多源信息融合问题［9］，因此可以将多尺度边缘测度的融合看作存在不确定和冲突的信息融合过程。
综上，本文提出一种基于证据推理的多尺度融合边缘提取算法。首先计算图像多尺度边缘测度响应，然后引入证据推理理论，将多尺度边缘测度作为证据推理的证据体，采用文中给出的双向指数法构造多尺度边缘测度的基本置信指派(Basic Belief Assignment, BBA)，然后利用冲突再分配DSmT组合规则（Conflict Redistribution DSmT，CR.DSmT）进行融合，通过双阈值方法确定边缘像素点并进行非极大值抑制和细化，最后通过对可见光和合成孔径雷达(Synthetic Aperture Radar,SAR)图像的边缘检测实验对算法的有效性进行了验证。
1 基于证据推理的鲁棒边缘检测算法
1.1 多尺度边缘检测算子提取
由于不同尺度下边缘测度会存在较大冲突，而Dempster理论在处理高冲突问题时，会出现Zadeh悖论和BBA分配不合理情况［10］，DSmT在多证据融合下，主焦元BBA难以收敛，因此采用新近提出的CR.DSmT组合规则［11］，该组合规则在DSmT辨识框架下，对冲突按一定原则再分配，解决Zadeh悖论的同时相对较好地处理了主焦元BBA收敛问题。假设辨识框架为2Θ，{θ1,θ2,θ3,θ4,θ5,…}为证据中的焦元，则：
m(φ)=0(8)
当组合的焦元中不包含冲突焦元时:
m(A)=∑X1,X2,…,Xk∈DΘX1∩X2∩…∩Xk=A∏ki=1mi(Xi) (9)
当冲突焦元之间相互组合以及冲突与不确定信息焦元组合时:
m(U)=∑X1,X2,…,Xk∈DΘUu(X1)∪u(X2)∪…∪u(Xk)∏ki=1mi(Xi)(10)
其中u(X）是组成X的所有θi的并集。
当冲突焦元与涉及产生该冲突的单焦元组合时:
m(U)=∑X1,X2,…,Xk∈DΘUX1∪X2∪…∪Xk∏ki=1mi(Xi)(11)
式(8)～(11)为CR.DSmT组合公式，将多尺度边缘测度BBA图像逐点采用CR.DSmT组合公式融合后，保留融合后的BBA图像。 (责任编辑：南粤论文中心)转贴于南粤论文中心: http://www.nylw.net(南粤论文中心__代写代发论文_毕业论文带写_广州职称论文代发_广州论文网）
- 毕业论文定制
南粤论文中心提供本科/研究生/博士毕业论文，从提纲，开题到后期免费修改,免费指导答辩等 ，联系人：魏老师，电话：13450375436。 - 代写代发职称论文
南粤论文中心提供职称论文代写，报告总结，演讲致辞,心得体会,党团辅导等代写服务并推荐发表省级,国家级,核心杂志包发。联系QQ:631069747。
基于冲突再分配DSmT的多尺度融合边缘
来源：网络(转载) 作者：乔奎贤 伊诗白 曲圣 发表于：2012-04-17 11:14 点击：次
【关健词】边缘检测；冲突再分配；多尺度边缘测度；基本置信指派；证据理论
由于实际景象地物特征复杂，单一尺度边缘检测算子提取的边缘与噪声点测度差异小，因此将导致细小地物与噪声相互掺杂，边缘提取不准确的现象。针对此问题，提出了一种基于冲突再分配DSmT的多尺度融合边缘检测算法。首先提取图像多尺度边缘测度，接着提出双向指数映射基本
顶一下
(0)
0%
踩一下
(0)
0%
版权声明：因本文均来自于网络，如果有版权方面侵犯,请及时联系本站删除.
