CAJ | 학술논문

不同传感器获得的多幅图像对同一场景的描述具有信息的冗余性和互补性，在小波图像融合的基础上提出了一种基于边缘和方向相关性的图像融合算法。对参加融合的两幅图像进行小波多尺度分解，对高低频子带系数分别采取不同的融合方法；根据高频分量不同的小波方向特性，采用基于方向相关系数的融合规则，而对低频分量采用能量和边缘特征相结合的融合方法；通过小波重构得到融合图像。实验结果表明，融合图像的视觉效果较好，客观评价指标均有提高。
불동전감기획득적다폭도상대동일장경적묘술구유신식적용여성화호보성，재소파도상융합적기출상제출료일충기우변연화방향상관성적도상융합산법。대삼가융합적량폭도상진행소파다척도분해，대고저빈자대계수분별채취불동적융합방법；근거고빈분량불동적소파방향특성，채용기우방향상관계수적융합규칙，이대저빈분량채용능량화변연특정상결합적융합방법；통과소파중구득도융합도상。실험결과표명，융합도상적시각효과교호，객관평개지표균유제고。
A new image fusion rule based on edge characteristics and directional correlation was proposed to fuse the multiple images that there is some redundancy and complementarity between different images of the same scenarios .The wavelet multi-scale decomposition technique is applied to decompose these images that would be fused first .Then an algorithm was used to calculate the fusing coefficient of high fre-quency subband that is different from the algorithm for low frequency subband .The directional co-relation was proposed to calculate the fusing coefficient for every coefficient of the high-frequency in different di-rections,and the low-frequency approximation component was determined by image edge characteristics and energy .Final fusion image is obtained by wavelet reconstruction .The good visual effect of the simula-tion experimental results is obtained .All various indicators of objective evaluation index is increased .This method is proved to be effective .