汽车类外观专利图像三维重建  

3D reconstruction on automobile design patent images

作　　者：张丹[1] 吕学强[1] 李卓[1] 徐丽萍[2] 

机构地区：[1]北京信息科技大学网络文化与数字传播北京市重点实验室,北京100101 [2]北京城市系统工程研究中心,北京100089

出　　处：《计算机工程与设计》2014年第9期3136-3142,共7页Computer Engineering and Design

基　　金：国家自然科学基金项目(61171159;61271304);北京市教委科技发展计划重点项目暨北京市自然科学基金B类重点基金项目(KZ201311232037)

摘　　要：为实现由汽车类外观设计专利图像到三维模型的恢复,结合汽车类外观专利图像的特点,提出融合从明暗恢复形状(SFS)与轮廓线法的汽车类外观专利图像三维重建方法。结合右视图及主视图的轮廓对由SFS得到的模型进行深度校正,得到较为精确的单幅图像三维模型;利用轮廓法恢复物体的整体模型,通过计算各单幅图像模型及整体模型的位置坐标,进行各模型之间的对齐操作,使两种方法得到的模型融为一体,得到融合模型。实验结果表明,与其它方法相比,该方法得到的三维模型在整体与细节方面都得到较大提升。To recover the 3D model from automobile design patent images, a method that fused the SFS-method and the contourmethod combining the characteristics of the automobile design patent images was proposed. Contours of the right-view and the front-view were combined to correct the depth of the model obtained by using SFS and a more precise 3D model of a single image was recovered. The contour-method was used to recover the whole-model and by calculating the location of the single-model and the whole-model, the models were calibrated, so that the models obtained by using the two methods were fused successfully, and the fusion model was obtained. Experimental results show that, compared with other methods, not only the entirety aspect but also the details of the 3D model obtained by using the method are greatly improved.

关 键 词：汽车类外观专利图像 三维重建 从明暗恢复形状 轮廓线法 深度校正 融合模型 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]