纹理分区相融合的建筑区三维影像数据提取  

3D Image Data Extraction of Building Area Based on Fusion of Texture Partition

作　　者：林朝飞[1] 

机构地区：[1]攀枝花学院,四川攀枝花617000

出　　处：《科技通报》2015年第4期88-90,共3页Bulletin of Science and Technology

摘　　要：在城市筑群空间分布模式中,需要对建筑区的三维影像特征进行数据提取,用遥感影像中包含的光谱特征和纹理特征,实现对建筑物的特征融合检测。传统方法采用高亮线条和阴影区域融合的方法实现对建筑区的三维影像数据提取,当图像本身受斑点噪声影响较大和光程较远的情况下,影像数据提取效果不好。提出一种基于形状特征与纹理分区相融合的建筑区三维影像数据提取方法,采用骨架线模式和三角网模式构建建筑物的三维纹理模型,设计基于纹理分区的建筑区三维影像图像分割算法。在三角网纹理分区结构模型的基础上,通过相互近似正交的直线模式建立建筑群直线模式的方向关系图,实现对建筑区的三维影像数据提取。仿真实验表明,采用该模型对建筑区的三维影像数据提取,影像分割准确,能准确有效提取影像局部区域的亮度特征,受斑点噪声影响较大和光程的影响较小,对建筑区图像特征的检测正确率较高。In building group space distribution pattern of the city, need a three-dimensional image characteristic of building area for data extraction, using spectral features and texture features include remote sensing images, to realize the fusion detection characteristics of the building. The traditional method is using the highlight line method of realizing integration and shadow region extraction of 3D image data of the construction area, when the image itself affected by speckle noises and a larger path a distance, the effect of extraction of image data is not good. A 3D image data extraction method of building area is proposed based on fusion of shape feature and texture partition, construction of three-dimensional texture model building using skeleton line model and triangular mesh model, design of 3D image texture segmentation algorithm based on image partition construction zone. Based on triangular mesh texture partition structure model, establish the linear model by linear mode buildings approximately orthogonal direction diagram, to achieve the extraction of 3D image data of building area. Simulation results show that, using the extraction of 3D image data of the model for building area, image segmentation is accurate, it can accurately extract image brightness feature of local region affected by speckle noises, it has little effect of larger and optical path, detection rate of image feature construction area is high.

关 键 词：形状特征 纹理分区 建筑区 特征 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]