Since the modeling matching image has low definition in traditional landscape planning design based onvirtual per-spective,a high-density landscape planning design based on three-dimensional(3D)perspective is proposed. The binocular stereo vision method is used to convert the visual data,which can change the process from data scanning to feature extraction,and im-prove the validity of the data. The SIFT candidate feature extraction method is used to guarantee that the used data characteris-tics can represent the data attribute effectively,and improve the image identification degree and definition degree. In order to verify the validity of the design,the contrast simulation test was designed. The experimental results verify that the high-density landscape planning design based on 3D perspective can improve the definition of modeling matching image effectively,and has high reference value for the development of this field.%针对传统虚拟视角的景观规划设计中建模匹配图像清晰度不高的问题,提出一种基于三维视角的高密度景观规划设计.采用双目立体视觉方法进行视觉数据的转换,改变了传统意义上的扫描数据到特征提取的过程,提高了数据的有效性,同时使用SIFT候选特征提取方法,保证使用的数据特征能够有效地代表数据的属性,提高了图像的辨识度以及清晰程度.为了验证设计的有效性,设计了对比仿真试验,通过试验结果能够证明基于三维视角的高密度景观规划设计有效地解决了建模匹配图像清晰度不高的问题,对该领域的发展具有很高的借鉴价值.
展开▼
