古老建筑三维景观特征整合仿真研究

发布时间：2018-05-11 19:27

本文选题：古老建筑 + 三维景观　；参考：《计算机仿真》2017年11期

【摘要】：对古建筑三维景观特征的整合,能够有效在保护古老建筑的基础上进行现代化建设。对建筑三维景观特征的整合,需要对特征向量集进行加权处理,计算三维景观特征的概率分配函数。传统方法先得到三维景观纹理和颜色特征,利用在线计算来调整特征加权系数。但忽略了计算三维景观特征的概率分配函数,导致整合精度偏低。提出基于D-S证据理论的古老建筑三维景观特征整合方法,首先组建张量子空间,构造三维景观每个模上的协方差矩阵;其次计算三维景观特征向量和最优投影矩阵,对特征向量集进行加权处理,获得古老建筑三维景观的特征;然后采用负自然指数函数计算三维景观特征的概率分配函数和似然函数;最后采用D-S证据理论正交和公式对三维景观的特征进行整合。仿真结果证明,所提方法提高了整合图像空间质量,且保留了图像的光谱特性。
[Abstract]:The integration of three-dimensional landscape features of ancient buildings can effectively protect the ancient buildings on the basis of modernization. For the integration of architectural 3D landscape features, the eigenvector set needs to be weighted to calculate the probability distribution function of 3D landscape features. The traditional method first obtains 3D landscape texture and color features, and adjusts the feature weighting coefficient by on-line calculation. However, the probability distribution function of 3D landscape features is neglected, which leads to low integration accuracy. Based on D-S evidence theory, an integration method of 3D landscape features of ancient buildings is proposed. Firstly, a quantum space is constructed to construct the covariance matrix on each model of 3D landscape, and then the eigenvector and optimal projection matrix of 3D landscape are calculated. The feature vector set is weighted to obtain the features of 3D landscape of ancient buildings, and then the probability distribution function and likelihood function of 3D landscape feature are calculated by negative natural exponential function. Finally, D-S evidence theory is used to integrate the features of three-dimensional landscape by orthogonal summation formula. The simulation results show that the proposed method improves the spatial quality of the integrated image and preserves the spectral characteristics of the image.
【作者单位】：西安工程大学艺术工程学院;
【分类号】：TP391.41;TU-87;TU986

【相似文献】