Cross-correlation and image alignment for multi-band IR sensors

被引：0

作者：

Lu, Thomas ^{[1
]}

Chao, Tien-Hsin ^{[1
]}

Chen, Kang ^{[2
]}

Luong, Andrew ^{[3
]}

Dewees, Mallory ^{[4
]}

Yan, Xinyi ^{[5
]}

Chow, Edward ^{[1
]}

Torres, Gilbert ^{[6
]}

机构：

[1] CALTECH, Jet Prop Lab, Pasadena, CA 91125 USA

[2] Univ Calif Los Angeles, Los Angeles, CA USA

[3] Univ Calif Irvine, Irvine, CA USA

[4] Saddleback Coll, Mission Viejo, CA USA

[5] Univ Calif Berkeley, Berkeley, CA 94720 USA

[6] Naval Air Warfare Ctr, Point Mugu Nawc, CA USA

来源：

OPTICAL PATTERN RECOGNITION XXVII | 2016年 / 9845卷

关键词：

Multi-band IR images; cross-correlation; image processing; feature extraction; Wavelets;

D O I：

10.1117/12.2224694

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

We present the development of a cross-correlation algorithm for correlating objects in the long wave, mid wave and short wave Infrared sensor arrays. The goal is to align the images in the multi sensor suite by correlating multiple key features in the images. Due to the wavelength differences, the object appears very differently in the sensor images even the sensors focus on the same object In order to perform accurate correlation of the same object in the multi-band images, we perform image processing on the images so that the features of the object become similar to each other. Fourier domain band pass filters are used to enhance the images. Mexican Hat and Gaussian Derivative Wavelets are used to further enhance the features of the object. A Python based QT graphical user interface has been implemented to carry out the process. We show reliable results of the cross-correlation of the objects in multiple band videos.

引用

页数：12