Abstract: Large numbers of 3D model have been used in the product design and 3D artifact, but problems have occurred that attribute data is incomplete; the search method is inaccurate based on attributes. Consequently, the search engines can not provide effective search service. This paper proposes an area distributions based method. According to this method, summarize the total area and average area of the vertex of the 3D model, normalize the list of the area distributions list and Fourier transform the list; get the final area distributions list model, and map the search of the model to compare the area distributions list. Experiments have been conducted for comparing and evaluating the proposed algorithm utilizing the Engineering Shape Benchmark（ESB）database. The experiential results show that the proposed technique has effectively reflected the similarity among engineering models, and the match result of the extremely similar model is accurate and the retrieval performance has been significantly improved compared with traditional shape distribution method.

Key words: engineering model retrieval, area distributions, Fourier transform

摘要： 在产品设计、3D制作等领域中经常使用大量的三维模型，但现有的基于属性的检索方式由于属性数据不全、不准的问题，搜索引擎还不提供有效的三维模型检索服务。提出了一种与CAD系统无关的基于面积分布算子的三维模型检索算法，该算法采用模型中顶点关联的三角形面积总和及平均面积构成的面积序列，对面积分布序列进行归一化操作和傅里叶变换等处理，得到最终的面积分布算子，将三维模型的相似度计算映射为面积分布序列的比较。在普渡大学建立的工程标准模型库（ESB）上对该算法进行了验证和比较。实验结果表明该算法客观反映了模型的相似程度，明显提高了传统的检索算法的效率和速度。

关键词: 三维模型检索, 面积分布算子, 傅里叶变换