摘要
针对传统破损数据交互方法未考虑位置系数导致交互效果差的问题,提出基于双边非局部均值算法的视觉传感网络图像破损数据交互方法。根据采集图像与真实场景之间的关系,构建像素多尺度域模型和多传感器矢量模型,确定传感器噪声特征信息;引入双边非局部均值算法获得位置系数,结合像素及其周围像素点之间的关联性进行去噪处理;按照贝叶斯理论建立破损数据观测模型,计算图像破损区域分布函数的均值和方差;设定扩展结构,定义像素参数,提取图像梯度边缘特征;将归一化互信息作为交互准则,选择两幅图像做空间变换,并使用金字塔分解方法优化交互过程,减少计算量。仿真结果证明,所提交互方法能够达到图像修复、增强图像质量的目的。
Aiming at the problem that the traditional damage data interaction method does not consider the position coefficient, which leads to poor interaction effect, this paper proposes an image damage data interaction method based on the bilateral nonlocal mean algorithm. According to the relationship between the acquired image and the real scene, the pixel multi-scale domain model and multi-sensor vector model were constructed to determine the sensor noise feature information;the bilateral nonlocal mean algorithm was introduced to obtain the position coefficient, and the denoising was carried out combined with the correlation between the pixel and its surrounding pixels;the damage data observation model was established according to the Bayes theory. Firstly, the mean and variance of the domain distribution function of the image damage area were calculated;secondly, the extended structure was set, the pixel parameters were defined, and the gradient edge features of the image were extracted;thirdly, the normalized mutual information was used as the interaction criterion, two images were selected for spatial transformation, and the pyramid decomposition method was used to optimize the interaction process, so as to reduce the amount of calculation. Simulation results show that the proposed method can achieve the purpose of image restoration and image quality enhancement.
作者
陈纾
孟刚
CHEN Shu;MENG Gang(Fuzhou University,Fuzhou Fujian 350100,China;Nanjing Tech University,Nanjing Jiangsu 211816,China)
出处
《计算机仿真》
北大核心
2021年第10期185-188,203,共5页
Computer Simulation
关键词
视觉传感器
图像破损
数据交互
互信息
金字塔分解
Vision sensor
Image breakage
Data interaction
Mutual information
Pyramid decomposition
