基于局部结构形态改进图像边缘限幅滤波算法研究

来源 :计量学报 | 被引量 : 0次 | 上传用户:zzggwd
下载到本地 , 更方便阅读
声明 : 本文档内容版权归属内容提供方 , 如果您对本文有版权争议 , 可与客服联系进行内容授权或下架
论文部分内容阅读
应用经典限幅滤波算法(CFA)对边缘去噪处理时容易造成有效的高频边缘被抑制、破坏边缘连续性、丢失目标结构特征的问题.提出了一种基于局部结构形态特征的改进型限幅滤波算法,用于图像边缘滤波处理.该算法利用限幅滤波的原理,引入由5个相邻边缘点构成的滑动模子并遍历边缘各点,对滑动模子中增量超限的点加以结构形态预测和阈值判断,即通过滑动模子建立局部轮廓的结构形态模型,并应用模型进行边缘预测;对超限点与预测值的差异进行了比较,为判定是否遇到台阶、凸缘或尖锐的结构特征提供了依据.为了测试新算法在边缘保持和滤波降噪方面的能力,与传统限幅滤波算法进行了对比实验.实验结果表明:基于局部结构形态特征的改进限幅滤波算法不但具有高效的去噪能力,而且对目标结构中的高频边缘具有显著保护作用.
其他文献
The octahedral molecular sieve (OMS-2)-supported Fe (xFe/OMS-2:x =1,3,5,and 10) cat-alysts were prepared using the pre-incorporation method.Physicochemical properties of the as-synthesized materials were characterized by means of various techniques,and th
Triphenylmethane (tpm) derivatives (e.g.tpmcv) have threatened the safety of the aquatic environment due to the potential toxicity and carcinogenicity.In this study,the novel ul-trasonic/persulfate/chlorite (US/S2O82-/ClO2-) oxidation process was develope
Nanosized Ba1-xBixFe1-xCuxO3 (12-50 nm) with x values of 0,0.01,0.05,and 0.1 system was prepared using the Pechini method.Structural,morphological,surface and optical charac-terizations were performed for the prepared samples.Cubic phase was the predomina
The rapid construction of artificial reservoirs in metropolises has promoted the emergence of city-river-reservoir systems worldwide.This study investigated the environmental be-haviors and risks of heavy metals in the aquatic environment of a typical sys
We measured the concentrations of Cr,Fe,Ni,Cu,Zn,Cd,Pb and Hg,and the stable iso-tope ratios (δ13C and δ15N) in 87 fish samples within 12 economic fish species collected from the Dongting Lake,the second largest freshwater lake in China.With few exception
讨论了SI“秒”在量子基准中的重要性和坐标时在时空统一及其他物理量基准统一中的作用,以史瓦西场为例说明了“原时”与“坐标时”的关系以及引力场对“原时”的影响,认为时间单位秒的不确定度决定了其他物理量单位的不确定度.只有“坐标时”的适用范围是全局的,涉及到大尺度时空的物理量必须用也只能用坐标量进行表征.SI“秒”定义的是一个“原时”秒长,是一个局域物理量,要使用量子基准对大尺度空间的物理量进行量值统一,需要进一步明确“坐标时”与SI“秒”之间的关系.
Recovery of phosphorus (P) from wastewater is of great significance for alleviating the short-age of P resources.At present,the P recovery process is faced with the problem of excessive organic carbon consumption when obtaining a P-concentrated recovery s
Nanoparticles (NPs) are widely used for their special physical properties and released into the natural environment.When two types of NPs exist in the same environment,the pres-ence of one type of NP may affect the properties of the other type of NP.This
针对3D打印点阵结构中缺陷目标因尺寸小、缺陷特征微弱而难以准确自动识别的问题,提出了一种基于YOLOv3算法的点阵结构缺陷智能识别新方法.该方法利用深度学习网络模型在特征提取方面的优势,采用多尺度网络进行预测,将缺陷的分类和定位问题作为回归问题处理.实验结果表明,所提算法实现了一种3D打印点阵结构内部典型缺陷的识别,缺陷检测回召率为96.6%,准确率为93.2%,模型平均精度均值为0.957,为进一步精确表征缺陷并分析缺陷对点阵结构性能的影响提供了依据.
提出了一种基于包容直线评定不连续平面的方法.首先,基于主成分分析对点云数据进行全局噪声剔除;然后,采用最小二乘平面拟合算法、多阈值点云提取算法及平行直线拟合算法获取符合条件的包容直线,并以空间中2条平行的包容直线代替包容平面建立最小区域;最后实现不连续平面的平面度误差评定及指导平面调整.实验结果表明:该方法与传统的特征点法相比平面度误差值比遗传算法降低0.011 mm,并且根据输出的偏距信息对平面进行调整后,平面度误差值减小1.959mm.能有效地评定不连续平面的平面度误差和指导平面调整.