一种基于视图和支持向量机的三维物体识别方法

为提高三维物体识别系统性能并减少计算复杂性,本文提出了一种基于视图的方法.首先从三维物体的二维视图中提取颜色矩、纹理特征和仿射不变矩.颜色矩对于物体的大小和姿态不敏感且性能稳健.纹理特征可区别形状相似但外观不同的物体.仿射不变矩在物体发生仿射形变下具有不变性.本文将上述各种特征组合为23个分量的特征向量,送入支持向量机进行训练并识别.基于两种公开的三维物体数据库COIL-100和ALOI测试了本文方法性能.当每物体训练视角为36个(视角间隔10°)时,在两个数据库上的实验都达到了100%的识别率.进一步减少训练视角数量也达到较满意的识别性能,优于文献中的方法.

View-based Three Dimensional Object Recognition Approach Using Support Vector Machine

To improve the performance of three-dimeusional object recognition system and reduce computational complexity, a view-based method is proposed. First we extract color moments, texture features and affine invariant moments from the 2D view images of 3D objects. Color moments are robust and insensitive to the size and pose of objects. Texture features can distinguish objects which have similar shapes and different appearance. Affine invariant moments have the invariant properties under affine transformation. These features are combined into a feature vector of 23 elements, and then fed to Support Vector Machine (SVM) for training and recognition. We assessed our method based on two public 3D object dataset" COIL-100 and ALOI. 100% correct rate of recognition was obtained on both dataset when the number of presented training views for each object was 36 (10 degrees interval). When the number of training views was reduced, the correct rate of recognition was also satisfied and outperformed previous algorithms.