| 摘 要: 使用基于RANSAC提纯的改进SIFT算子和SIFT算子分别在两张不同视角的图像中提取并匹配特征点 对，通过实验结果验证了基于改进的SIFT算提取特征点的准确率和匹配的精度都要比SIFT算子和Harris算子高，改进 SIFT算子提取匹配出的特征点对更加的有利于目标物体的三维重建工作。 |
| 关键词: 物体特征提取和识别；SIFT；RANSAC；三维重建 |
|
中图分类号: TP391
文献标识码: A
|
|
| Research on Object Feature Extraction and Matching Based on Improved SIFT and RANSAC |
|
ZHANG Chunlin,CHEN Jinjie
|
( University of Shanghai for Science and Technoloy, Shanghai 200093, China)
|
| Abstract: The SIFT operator and the improved SIFT operator based on RANSAC purification are used to extract and match feature point pairs in two images with different perspectives.The experimental results verify that the accuracy and matching of feature points extracted based on the improved SIFT operator are higher than those of the SIFT operator and Harris operator.The feature points extracted and matched by means of the improved SIFT operator are more conducive to the 3D reconstruction of target objects. |
| Keywords: object feature extraction and matching;SIFT;RANSAC;3D reconstruction |
