立体匹配SIFT图像特征点提取算法仿真研究

Stereo Matching SIFT Image Feature Extraction Algorithm

研究图像特征优化提取技术。针对同一场景下的投影图像总是存在着一定的差异,造成了图像特征点提取精度低等不足,提出了一种改进的SIFT图像特征点提取算法,采用了SIFT算子算法和立体匹配算法相结合的混合算法。首先给出了图像特征点的相关定义,然后算法对所有像素点进行一次全面地判断,快速而准确的去除掉一部分没有关系的点,在剩余点中选取图像的特征点。实验结果证明了提出的方法能够有效地提取图像关键点信息。

The main method of image feature extraction technology was studied. For the same scene, certain differences always exist in the projection images, resulting in low image feature extraction accuracy. This paper pro- posed an improved SIFT image feature extraction algorithm, which uses the SIFT operator algorithm and stereo matc- hing algorithm hybrid algorithm. Firstly, the image feature points were defined, and then all the pixels were compre- hensively judged and the unrelevant points were removed quickly and accurately. In the remaining points, the image feature points were selected . The experimental results show that the proposed method can effectively extract the key image point information.