基于云分形的自然景物模拟研究  

Research on Natural Scene Simulation Based on Cloud Fractal

作　　者：张盼盼 秦昆[2] 郭培煌 ZHANG Panpan;QIN Kun;GUO Peihuang(Beijing Huawei Digital Technology Co.,Ltd.,Beijing 100095,China;School of Remote Sensing and Information Engineering,Wuhan University,Wuhan 430072,China;Beijing Horizon Information Technology Co.,Ltd.,Beijing 100094,China)

机构地区：[1]北京华为数字技术有限公司,北京100095 [2]武汉大学遥感信息工程学院,湖北武汉430072 [3]北京地平线信息技术有限公司,北京100094

出　　处：《地理空间信息》2022年第12期17-22,共6页Geospatial Information

基　　金：国家自然科学基金资助项目(42171448)。

摘　　要：分形理论在自然景物模拟上已经存在较为广泛的研究和应用,目前大部分研究都聚焦在物体的自相似性层面上,为了更好地体现自然景物模拟的真实性和逼真度,分析了各种常用的分形模拟方法和云分形方法,利用云模型特有的不确定性对传统分形算法进行优化,在自然景物的细节描述上引入随机性,尝试探索云模型在优化分形算法(如BSP+L系统、IFS方法)方面的有效性。实验结果表明,云分形方法在自然景物模拟中所体现的随机性、不确定性优势显著,细节描述多样性,呈现的自然景物更加逼真。Fractal theory on the natural scenery simulation has been widely research and application,at present most of the research is focused on the self-similarity inside object-level.To better reflect the authenticity and fidelity of the natural scenery simulation,we analyzed the various commonly used fractal simulation method and cloud fractal method,optimized traditional fractal algorithms with cloud model which had the uncertainty of characteristic features,introduced randomness into the detail description of natural scenes,and explored the effectiveness of cloud model in optimizing fractal algorithms(such as BSP+L system and IFS method).The experimental results show that the cloud fractal method has obvious advantages of randomness and uncertainty in natural scene simulation,the detail description is diversified,and the natural scene is more realistic.

关 键 词：分形理论 云模型理论 云分形 自然景物模拟 

分 类 号：P237[天文地球—摄影测量与遥感]