结合三维几何形状信息和二维纹理的3D目标匹配

3D object matching combined 3D geometrical shape and 2D texture feature

为了实现场景中三维目标与模型之间的匹配,提出了一种结合三维几何形状信息和二维纹理的三维目标匹配方法。首先提取场景中深度图像的尺度不变特征变换(SIFT)特征,用SIFT算法与三维模型重建时所用到的一系列2.5维深度图像进行一一匹配,找到与场景中目标姿态最为相似的深度图像,提取此深度图像的三维几何形状特征与模型进行匹配,实现模型的初始化,即将模型重置到与场景目标相接近的姿态。最后用融合二维纹理信息的迭代就近点(ICP)算法实现场景中目标与模型之间的匹配,从而得到场景中三维目标的准确姿态。实验结果验证了方法的可行性与精确性。

To solve the matching problem between the model and 3D object in the scenes, this paper presented a 3D object matching method combined 3D shape and 2D texture feature. Scale-Invariant Feature Transform （SIFT） feature was extracted from the range image in the scene, and then the range image matched with a series of 2.5 dimensional range images which were used for the 3D model reconstruction one by one based on SIFT algorithm, so that it could find out the most similar local range image to the object in the scene. The matching between this local range image and the object was completed through 3D shape feature. It is to initialize the model, in other words, it is to reset the model close to the object in the scene. At last, a herative Closest Point （ICP） algorithm combined with color was used to implement the matching between the object in the sences and the model which was reset before. In this way the pose of the object in the scene can be calculated accurately. The experimental results verify the feasibility and accuracy of the proposed method.