机器视觉中特征点提取算法的探讨

A Discussion of Interest Operators for Extracting Point-Features in Computer Vision

机器视觉中特征点的提取是基于点特征图像配准算法的一个非常关键的步骤。首先从特征点定义出发,介绍了近年来国内外出现的几种常用的点特征提取算法,然后利用特征提取方法评价标准,对各种检测算子进行了综合的分析和讨论,最后,通过运用VC++编程语言工具进行算法仿真,对试验结果进行比较分析,理论与实践相结合,合理评价了各种方法的性能、优缺点、适用环境,论述对学习图像检测和具体工程应用具有很好的参考价值,并且为这些方法的改进提供了指导性意见。

Feature point extraction is a key step based on the point - feature characteristics of image registration, some common point- feature extraction algorithms were introduced first in this paper. And some evaluation criteria for feature extraction methods evaluation were used to perform comprehensive analysis and discussion. Then each al- gorithm was validated by VC ＋ ＋ with examples and the experimental results were compared and analyzed with theory and practice. These criteria can be used to effectively evaluate the performance, advantages and disadvantages of various feature extraction methods. These expositions have a good reference value for the learning of image detection and specific application, and provide direction for further improvement.