摘要
电路板红外图像配准技术是电路板红外热成像故障检测系统的核心。针对传统图像配准算法精度和效率较低等不足,为了提高电路板红外图像配准精度和效率,设计了一种基于ROI灰度压缩的图像配准算法。首先,利用Otsu法提取出电路板红外图像ROI。然后,将ROI内灰度值进行线性压缩。最后将灰度压缩后的图像进行互信息量运算和图像x平移、y平移、旋转、缩放变换参数寻优。仿真结果表明,基于ROI的灰度压缩算法能够显著减小配准时间,当两幅图像环境温度差异较大时其配准精度相比其他灰度压缩算法有明显优势。
Infrared image registration technique is the core of circuit board fault detection system. In order to improve the accuracy and efficiency of circuit board infrared image registration, an image registration algorithm based on ROI gray compression is proposed. First, Otsu method is used to extract circuit board infrared image ROI. Then, the ROI gray value is made linear compression. Finally the best image transform parameters are searched like x translation,y translation, rotation, scaling by mutual information calculation. The simulation results show that the image compression algorithm base on ROI can significantly reduce the registration time. The environment temperature difference of two images is larger, the registration accuracy is better.
出处
《激光与红外》
CAS
CSCD
北大核心
2014年第3期313-318,共6页
Laser & Infrared
基金
中央高校基本科研业务项目(No.ZXB2011A003)
中国民航大学科技基金项目(No.2010kyE07)
中国民航大学校级科研(No.2011kyE01)
中国民航机务维修科研基地资助
