基于复合模型的家居场景生成方法

发布时间：2018-06-19 11:53

本文选题：室内建模 + 自动布局　；参考：《计算机学报》2017年11期

【摘要】：室内三维场景生成技术近年来已得到广泛应用,但如何高效和自动化地生成场景元素的空间组织,仍然是个具有挑战性的问题.为了在给定的目标空间中快速生成符合日常生活规范的家居场景,提高室内场景建模的效率,该文基于室内设计领域的住宅空间家具配置流程,提出用于布局计算的复合模型,处理场景元素的组织和布局问题.具体地,该复合模型将室内场景中的对象分为代理区域和场景元素两个层次,首先以代理区域为单位,依据人体工程学的布局规则在目标空间中进行自动布局计算,然后通过案例库搜索,在代理区域中适配具体的场景元素,形成完整的家居场景.实验表明,与以家具为个体进行自动布局计算的方法相比,用代理区域代理一组家具进行布局减少了目标空间中参与布局的个体数目,降低了布局复杂度,在相同的运行时间下,能得到更为合理的布局结果.通过基于复合模型的家居场景生成方法,用户只需提供户型结构数据,而无需提前一一指定具体的家居元素,能给用户使用带来较大便利.
[Abstract]:Indoor 3D scene generation technology has been widely used in recent years, but how to efficiently and automatically generate spatial organization of scene elements is still a challenging problem. In order to quickly generate household scenes in a given target space and improve the efficiency of indoor scene modeling, this paper presents a composite model for layout calculation based on the furniture configuration flow of residential space in the field of interior design. Handle the organization and layout of scene elements. Specifically, the composite model divides the objects in the indoor scene into two levels: the agent area and the scene element. Firstly, the agent area is taken as the unit, and the automatic layout calculation is carried out in the target space according to the layout rules of ergonomics. Then through the case base search, the agent region adapts the specific scene elements, forms the complete home scene. The experimental results show that compared with the method of automatic layout calculation based on furniture, the number of individuals involved in the layout is reduced and the complexity of layout is reduced by using a group of furniture agents in the target space. Under the same running time, more reasonable layout results can be obtained. Through the method of household scene generation based on compound model, users only need to provide household structure data without specifying specific household elements in advance, which can bring great convenience to users.
【作者单位】：中国科学院计算技术研究所前瞻研究实验室;移动计算与新型终端北京市重点实验室;中国科学院大学;
【基金】：国家自然科学基金(61100086) 国家“八六三”高技术研究发展计划项目基金(2013AA013902)资助~~
【分类号】：TP391.9;TU241

【相似文献】