菱镁矿矿山井筒视频图像匹配方法的改进研究

Magnesite mine shaft improvement research of video image matching method

传统方法在菱镁矿矿山井筒视频图像的匹配过程中存在匹配精度差、误差率高等不足,不能满足视频图像匹配实时性的要求。提出一种基于优化的尺度不变特征转换算法,在保证图像配准精度的前提下,减少井筒视频图像的尺度变化,构造特征描述子并进行特征配准,实现对菱镁矿矿山井筒视频图像的精确匹配。

The traditional methods in magnesite mine shaft in the process of video image matching the matching accuracy and high error rate is insufficient, can't satisfy the requirement of matching real-time video images. Put forward a kind of scale invariant feature transform algorithm based on optimization, under the premise that guarantees the precision of image registration, reduce the change of the scale of the video image, wellbore structure feature descriptor and feature matching, implementation of magnesite mine shaft of video image exact match.