基于多小波包的红外/被动毫米波图像融合  

IR/PMMW Image Fusion Based on Multi-Wavelet Packet

作　　者：何伏春 聂建英[1] 

机构地区：[1]福州大学数学与计算机科学学院,福州350108

出　　处：《电光与控制》2015年第6期11-16,共6页Electronics Optics & Control

基　　金：福州大学科研基金(0030824649)

摘　　要：红外/被动毫米波(IR/PMMW)复合制导是当前发展多模复合制导技术的热点方向。红外探测系统在低能见度条件下穿透能力不如被动毫米波,而被动毫米波探测图像分辨率不如红外图像。为了更好地识别目标,提出一种新方法,首先采用多小波包与邻域平均相结合的去噪方法对IR/PMMW图像进行去噪处理;然后利用多小波包方法获取多张相似的IR/PMMW图像,再将它们进行主成分融合并进行精确的直方图规定化;最后,根据被动毫米波图像的特殊性,提出基于区域的图像质量评价方法。实验仿真结果表明,利用新方法处理图像,提高了图像的清晰度,更容易识别出目标信息。Infrared/passive millimeter wave （IR/PMMW） combined guidance is currently the hot trend in the field of multi-mode combined guidance technology. The penetrability of infrared detection system under low visibility conditions is not as good as that of passive millimeter wave, while the passive millimeter wave detection image is not as good as infrared image in resolution. In order to better identify the target, a new method was proposed. Firstly, IR/PMMW image was de-noised by using the de-noising method of combing multi-wavelet packet with neighborhood average, and then multiple similar IR/PMMW images were obtained by using the multi-wavelet packet method, which were then fused with principal component analysis, and accurate histogram specification was conducted. Finally, according to the particularity of PMMW image, a region based image quality evaluation method was proposed. The experimental simulation results indicate that after processing with the new method, the definition of image is improved, and it＇s easier to identify the target information.

关 键 词：被动毫米波图像 图像去噪 图像融合 多小波包 主成分分析 

分 类 号：V271.4[航空宇航科学与技术—飞行器设计] TN955[电子电信—信号与信息处理]