摘要
由于忽略了图像以真实性为准则的视觉传达需求,导致以形状基重构为主的相关方法效果不佳。为此,从视觉约束性出发,改进三维图像重构算法。分割、筛选所采集的目标点云数据,构建数据集。利用局部空间差分算法求解各像素值,并获得形状基约束性的像素值差分系数填充像素,以赋予各点云数据视觉传达约束性。以基于改进量子粒子群的k-means聚类算法和基于数字特征的二型熵模糊C均值聚类算法为基本算法,通过聚类目标区域和背景区域的数据,转换实际目标坐标系和重构图像坐标系,完成不同尺度、不同角度的三维图像重构。结果表明,所提方法重构的图像更贴合目标建筑的实际信息,符合图像的视觉传达需求,且峰值信噪比始终高于85 dB,均方误差均低于70 pixel,重构效果优越性显著。
Due to the neglect of the visual communication requirement of image authenticity as the criterion,the related methods with shape-based reconstruction have poor results.Therefore,the 3D image reconstruction algorithm is improved from visual constraints.This paper segments and filters the collected target point cloud data,and constructs a dataset.The local spatial difference algorithm is used to solve for each pixel value,and shape-based constraint pixel value difference coefficients are obtained to fill the pixels,which gives visual communication constraints to point cloud data.The k-means clustering algorithm based on improved quantum particle swarm and the type-2 entropy fuzzy C-means clustering algorithm based on digital features are used as the basic algorithms.By clustering the data of the target area and background area,an actual target coordinate system is transformed,and the reconstructed image coordinate system is reconstructed to achieve 3D image reconstruction at different scales and angles.The results show that the reconstructed images using the proposed method are more similar to the actual information of the target building,meet the visual communication needs of the images,and the peak signal-to-noise ratio is always above 85 dB,and the mean square error is below 70 pixel.The reconstruction effect is significantly superior.
作者
王晶
范晓鹏
WANG Jing;FAN Xiaopeng(Department of Culture and Media,Shaanxi Youth Vocational College,Xi’an 710068,China;Architecture College,Xi’an University of architecture and Technology,Xi’an 710055,China)
出处
《微型电脑应用》
2024年第9期178-181,共4页
Microcomputer Applications
关键词
视觉传达约束
三维图像
约束性赋予
聚类算法
坐标系转换
图像重构
visual communication constraint
3D image
constrain endowment
clustering algorithm
coordinate system transformation
image reconstruction
