全天候红外与可见光图像互信息最大化配准

(西安工业大学 电子信息工程学院,西安710021

红外; 可见光; Canny算子; 互信息

Mutual Information Maximization Registration of All-Weather Infrared and Visible Images
BAI Hao,GUO Quanmin,CHAI Gaixia

(School of Electronic Information Engineering,Xi'an Technological University,Xi'an 710021,China)

infrared; visible light; Canny operator; mutual information

DOI: 10.16185/j.jxatu.edu.cn.2019.06.004 http://xb.xatu.edu.cn

备注

为了配准两幅质量差异较大、像素相关性较小的异源传感器图像,采用直方图均衡技术增强待配准图像的质量,基于Canny算子提取质量增强后图像的边缘,利用互信息最大化算法计算了两幅边缘图像的配准矩阵,运用计算出的配准矩阵实现了质量增强后的原图像的配准。结果 表明:与直接采用互信息最大化算法配准的图像相比,采用本文方法配准图像的空间频率、标准差和平均梯度指标均有明显提高。

In order to register two heterologous sensor images with large quality difference and small pixel correlation,the histogram equalization technique is used to enhance the quality of the images to be registered.The edges of the enhanced images are extracted based on the Canny operator.The registration matrix of the two edge images is calculated by the mutual information maximization algorithm.The original image with the enhanced quality is registered by using the registration matrix obtained.The results show that the spatial frequency,standard deviation and average gradient of the image registered by this new method are significantly improved compared with the image registered directly by the mutual information maximization algorithm.