摘要
给出了一种基于统计形状模型的双目识别系统实现方法。该系统根据识别目标在场景中的姿态有一定规律性的特点,用先验的样本信息指导识别的过程,并采用径向对齐约束的方法对摄像头内外参数进行标定,利用图像识别结果中特征点的顺序性完成特征匹配,进而获得目标的三维信息。
The authors proposed a realization approach for binocular visual recognition system based on the active shape model. Considering the regularity of targets＇ pose in the scene, the system uses priorisamples＇ information to guide the process of recognition, and calibrates inner and outer parameters of two cameras under the radial alignment constraint. Then the system matches corresponding points by using sequential order of each feature point in recognition results of images, and finally obtains the 3D information of the target efficiently.
出处
《计算机应用与软件》
CSCD
2009年第3期223-225,共3页
Computer Applications and Software
关键词
双目识别
主动形状模型
径向对齐约束
三维重建
Binocular visual recognition Active shape model Radial alignment constraint 3D reconstruction
