摘要
将深度图像和灰度图像相结合,对围绕定轴旋转的三维目标进行了识别与分类。将深度图像作为相位因子,对其进行傅里叶变换,并用其制成三维定向图,用于三维目标的识别和旋转角度的判定;对于灰度图像,采用主分量分析(PCA)的方法,对训练图像进行特征分析。根据深度图像测定的目标角度,对三维目标灰度图像在其所属特征空间进行分解与重构。实验结果表明,综合利用深度图像和灰度图像,可以大大降低目标识别中的误判概率。
The range image and the texture image are combined to recognize and classify the three-dimensional（3D） objects which rotate around the fixed axis.The range image is considered as a phase factor,and then the phase Fourier transform of the factor is used to build a 3D object orientation map,which is used for 3D objects identification and to detect the rotation angle;The texture images are processed by principal component analysis（PCA）,and the feature vectors are selected.According to the detected angle,we choose the corresponding feature space to decompose and reconstruct the texture image of 3D object.The results indicate that the proposed scheme can reduce the error rate of object recognition obviously.
出处
《光电子．激光》
EI
CAS
CSCD
北大核心
2010年第2期312-317,共6页
Journal of Optoelectronics·Laser
基金
江苏省高校自然科学研究重大项目(09KJA140002)
江苏省自然科学基金资助项目(BK2006726)
