CAJ | 학술논문

针对红外与可见光图像的特点和Contourlet变换的多分辨率和多方向的特性，提出了一种基于Contourlet变换的红外与可见光图像融合方法。该方法将源图像做Contourlet变换，获得低频分量和各带通方向子带的高频分量，根据低频和高频分量各自的特点采用不同融合规则。低频分量反映源图像的平均特性，包含源图像的光谱信息和大部分的能量信息，对低频分量采用基于块的标准差加权的融合规则；针对高频分量包含大量有用的边缘和纹理等细节信息，对融合图像质量影响较大，考虑到不同尺度上的高频系数包含的信息量不同，对高频分量的不同尺度的系数制定不同的融合规则，在较大尺度上采用系数绝对值取大的方法，在较小尺度上采用小系数区域能量取大的方法和大系数绝对值取大的方法。最后经Contourlet逆变换重构获得最后的融合图像。仿真结果表明，通过主观视觉效果和客观评价指标进行评价，与传统图像融合算法相比，该方法能有效提取源图像的有用信息，取得较好的融合图像结果。
침대홍외여가견광도상적특점화Contourlet변환적다분변솔화다방향적특성，제출료일충기우Contourlet변환적홍외여가견광도상융합방법。해방법장원도상주Contourlet변환，획득저빈분량화각대통방향자대적고빈분량，근거저빈화고빈분량각자적특점채용불동융합규칙。저빈분량반영원도상적평균특성，포함원도상적광보신식화대부분적능량신식，대저빈분량채용기우괴적표준차가권적융합규칙；침대고빈분량포함대량유용적변연화문리등세절신식，대융합도상질량영향교대，고필도불동척도상적고빈계수포함적신식량불동，대고빈분량적불동척도적계수제정불동적융합규칙，재교대척도상채용계수절대치취대적방법，재교소척도상채용소계수구역능량취대적방법화대계수절대치취대적방법。최후경Contourlet역변환중구획득최후적융합도상。방진결과표명，통과주관시각효과화객관평개지표진행평개，여전통도상융합산법상비，해방법능유효제취원도상적유용신식，취득교호적융합도상결과。
Aiming at the features of infrared and visible images and the advantages of multi?scale and multiple directions in the Contourlet transform,this paper proposes a method for fusion of infrared and visible images based on Contourlet transform.Firstly,the original images are decomposed by using Contourlet transform to obtain low and high frequency coefficients.The different fusion rules are used based on the different features of low and high frequency coefficients.For low frequency coefficients,the weighted regional standard variance fusion rule is adopted in view of the image average characteristics.For high frequency coefficients,and a fusion rule based on multi?scale is adopted in the view of different scale with different information. The maximum absolute coefficients are used with bigger scales.The rest scales are divided into two parts:small and big coefficients.This paper chooses the biggest absolute coefficient with big coefficients,and adopts the coefficients with maximum regional power of small coefficients. Finally, the final fusion image is acquired through Contourlet inverse transform.The simulation results show that the proposed method is superior both in vision and in objective evaluation compared with traditional image fusion methods,and can get more useful information from original images and better image fusion.