一种电力设备红外与可见光配准研究

A Registration Algorithm for Infrared and Visible Images of Power Equipment

针对电力设备红外与可见光图像配准因源图像模糊、存在视角差异等问题,导致配准精度低的缺点,提出一种基于轮廓特征相红外与可见光图像配准方法。首先以轮廓中线特征主方向为检测到的特征点分配方向。然后使用连线一致性匹配方法,筛选出正确的特征点。最后得到配准结果。实验结果表明,该算法增强了红外与可见光源图像清晰度,解决了视角差异问题,提高了图像配准的精度。

In view of the low registration accuracy caused by the blurred source image and the difference of perspective in the infrared and visible light image registration of power equipment,a method of infrared and visible light image registration based on contour feature phase is proposed.Firstly,the main direction of the contour midline feature is taken as the distribution direction of the detected feature points.Then the correct feature points are selected by using the line consistency matching method.Finally,the registration results are obtained.The experimental results show that the algorithm enhances the clarity of infrared and visible source images,solves the problem of angle difference and improves the accuracy of image registration.