共查询到20条相似文献,搜索用时 312 毫秒
1.
微观组织结构是判定焦炭质量的重要依据,通常需要偏光显微镜对焦炭光片成像,但显微镜头的特殊性,使得图像中间清晰而周围存在离焦模糊,降低了图像的质量,影响对焦炭组织成分的判定。因此,本文结合局部模糊分割算法和离焦去模糊网络,提出一种局部去模糊算法。首先对焦炭图像进行基于LBP的模糊检测,得到模糊度量图;通过Alpha抠图算法分割度量图,获得α模板矩阵;最后,以α模板矩阵作为融合比列,对原图和全处理图像进行线性融合,得到去模糊图像。通过实验对比,证明本文方法在累加概率(CPBD)等4个质量评价标准上面都得到了提升,相比于传统方法,去模糊效果更佳。 相似文献
2.
针对当前散焦模糊区域检测算法对于均质清晰区域容易误判,边缘定位不够准确的问题,提出一种基于LBP特征与图像显著性的散焦模糊区域检测算法。首先,利用LBP特征和SLIC算法来获取SLBP模糊图,利用DRFI显著性检测算法来获取DRFI显著图;然后,利用SLBP模糊图和DRFI显著图来构造三元标识图,进而利用KNN抠图算法来获取模糊图;最后,借助于形态学运算和平滑滤波来细化模糊图。在公共模糊数据集上的实验结果表明,该算法能有效地检测出均质清晰区域,保留图像边缘细节,在检测精度和查全率指标上表现较好。 相似文献
3.
4.
5.
6.
分块融合易产生块效应,而Contourlet变换融合方法不具备平移不变性.结合以上方法的各自优势,提出1种新的基于Contourlet变换与块分割相结合的多聚焦图像融合方法.首先对图像进行分块并评价子块清晰度,然后通过建立标记矩阵,确定图像清晰区域、模糊区域及两者的交界区域.对非交界区域,直接选取清晰子块的像素值作为融合后的像素值;对介于清晰与模糊之间的交界区域,采用1种新的基于Contourlet变换的融合算法完成融合.实验结果表明,该方法有效抑制了图像块效应,且具备平移不变性,融合效果明显优于分块和Contourlet变换融合方法. 相似文献
7.
8.
9.
结合粗糙集理论,利用像素邻域的空间信息,对图像色彩的分布进行了量化表示,并据此提出一种基于量化粗糙信息的改进图像分割方法,该方法使用局部量化粗糙度和待定算子来更新FCM算法中的隶属度函数,实现图像的初步分割,并针对初步分割后的小区域和相似区域,进行色彩区域的合并操作,来对分割结果进行优化。实验结果表明,相对于传统的模糊C-均值(FCM)聚类分割算法,提出的方法降低了时间复杂度,且具有良好的分割效果。 相似文献
10.
针对岩石薄片全幅面偏光采集中,部分视域采集的图像会存在离焦模糊问题.本文提出一种模糊图像检测方法,能在薄片全幅面采集的大量偏光图像中,自动检测出存在离焦模糊的视域.由于图像模糊会导致图像空间域和频率域的一些特征产生变化,因此本文结合空间域和频率域方法,对图像分块进行模糊评价得到模糊度图,由该图统计特性将图像区分为清晰图... 相似文献
11.
Jufeng Zhao Huajun Feng Zhihai Xu Qi Li Xiaoping Tao 《Signal, Image and Video Processing》2013,7(6):1173-1181
For images with partial blur such as local defocus or local motion, deconvolution with just a single point spread function surely could not restore the images correctly. Thus, restoration relying on blur region segmentation is developed widely. In this paper, we propose an automatic approach for blur region extraction. Firstly, the image is divided into patches. Then, the patches are marked by three blur features: gradient histogram span, local mean square error map, and maximum saturation. The combination of three measures is employed as the initialization of iterative image matting algorithm. At last, we separate the blurred and non-blurred region through the binarization of alpha matting map. Experiments with a set of natural images prove the advantage of our algorithm. 相似文献
12.
《Journal of Visual Communication and Image Representation》2014,25(5):1031-1043
In this paper, we propose a fully automatic image segmentation and matting approach with RGB-Depth (RGB-D) data based on iterative transductive learning. The algorithm consists of two key elements: robust hard segmentation for trimap generation, and iterative transductive learning based image matting. The hard segmentation step is formulated as a Maximum A Posterior (MAP) estimation problem, where we iteratively perform depth refinement and bi-layer classification to achieve optimal results. For image matting, we propose a transductive learning algorithm that iteratively adjusts the weights between the objective function and the constraints, overcoming common issues such as over-smoothness in existing methods. In addition, we present a new way to form the Laplacian matrix in transductive learning by ranking similarities of neighboring pixels, which is essential to efficient and accurate matting. Extensive experimental results are reported to demonstrate the state-of-the-art performance of our method both subjectively and quantitatively. 相似文献
13.
In this paper, we propose a single image deblurring algorithm to remove spatially variant defocus blur based on the estimated blur map. Firstly, we estimate the blur map from a single image by utilizing the edge information and K nearest neighbors (KNN) matting interpolation. Secondly, the local kernels are derived by segmenting the blur map according to the blur amount of local regions and image contours. Thirdly, we adopt a BM3D-based non-blind deconvolution algorithm to restore the latent image. Finally, ringing artifacts and noise are detected and removed, to obtain a high quality in-focus image. Experimental results on real defocus blurred images demonstrate that our proposed algorithm outperforms some state-of-the-art approaches. 相似文献
14.
2D图像转3D图像是解决3D影视内容缺乏的主要手段之一,而深度提取是其中的关键步骤.考虑到影视作品中存在大量散焦图像,提出单幅散焦图像深度估计的方法:首先通过高斯卷积将散焦图像转换成两幅模糊程度不同的图像;其次计算这两幅图像在边缘处的梯度幅值比例,进而根据阶跃信号与镜头的卷积模型得到边缘处的模糊度;再次将边缘处的模糊度转换成图像的稀疏深度并利用拉普拉斯矩阵插值得到稠密深度图;最后通过图像的视觉显著度提取前景对象,建立对象引导的深度图优化能量模型,使前景的深度趋于一致并平滑梯度较小区域的深度.该方法利用对象引导的深度优化,剔除了拉普拉斯矩阵插值引入深度图的纹理信息.模拟图像的峰值信噪比和真实图像的视觉对比均表明该算法比现有方法有较大改善. 相似文献
15.
边界模糊图像不同区域之间没有明确的分界,用传统的图像分割方法难以得到很好的分割结果。本文研究了径向基函数网络的工作实质及其用于图像分割的机理,分析了径向基函数神经网络的特点,针对边界模糊图像,应用不同结构的径向基函数神经网络对其进行图像分割,验证了径向基函数网络用于图像分割的有效性以及算法速度上的优越性。 相似文献
16.
基于支持向量回归的无参考模糊和噪声图像质量评价方法 总被引:6,自引:4,他引:2
基于支持向量回归(SVR)和图像奇异值分解,提出了一种新的无参考(NR,no-reference)模糊和噪声图像质量评价(IQA)方法。首先通过对待评价图像进行高斯低通滤波生成再模糊图像,然后分别对它们进行奇异值分解并计算奇异值的改变量,最后使用奇异值的改变量作为SVR的输入,训练预并测得到图像的质量评分。在3个公开的模糊和噪声数据库上的实验结果表明,新方法预测得分与主观得分有较好的一致性,获得了较好的评价指标;对于模糊失真类型和噪声失真类型,在LIVE2数据库上的性能评价指标斯皮尔曼等级相关系数(SROCC)分别达到0.961 3和0.965 9。 相似文献
17.
Wei Zhang Q.M. Jonathan Wu Guanghui Wang Xinge You Yongfang Wang 《Journal of Visual Communication and Image Representation》2010,21(4):271-282
Affine-invariant region detection is the basic technique for visual matching and has been widely applied in many areas. In this paper, we propose a simple yet effective method to detect the affine-invariant regions from gray image, which is called enclosed region. The enclosed region is detected based on the observation that one physical object is enclosed by the same region before and after affine transformation. The proposed method is a three-step method. Firstly, we segment the initial regions by using thresholds on the image. Secondly, external enclosing region (EER) and internal enclosed region (IER) are defined for each initial region, and we select the enclosed regions from the initial regions through applying histogram constraints on EER and IER. Thirdly, the largely overlapping regions are removed. Experiments on typical images exhibit the robustness of the proposed enclosed region detector. Extensively quantitative evaluation and comparison demonstrate that the proposed method outperforms state-of-the-art methods. 相似文献
18.
19.
为了准确有效地实现自然图像的压缩感知重构,提出一种使用拉普拉斯尺度混合(Laplacian Scale Mixture,LSM)先验的结构化近似消息传递(Approximate Message Passing,AMP)算法.利用LSM模型构建AMP算法的高阶统计约束,将压缩感知重构问题转化为先验信息估计问题和奇异值最小化问题.首先,用LSM分布刻画相似块矩阵奇异值的稀疏性,其中该稀疏性指示了图像块的相似性,因此LSM模型被用来描述图像的非局部相似结构;然后,通过期望最大化算法估计LSM模型的尺度参数,得到可靠的先验信息;最后,由AMP算法求解奇异值最小化问题,实现图像的精确重构.实验结果表明,提出的结构化AMP算法的图像重构质量优于多种主流的压缩感知图像重构算法. 相似文献