CAJ | 학술논문

为了提升红外与可见光图像融合精度，提出了一种局部化非下抽样剪切波变换与脉冲耦合神经网络相结合的红外与可见光图像融合方法。首先，利用局部化非下抽样剪切波对源图像进行多尺度、多方向分解；然后，在分解后的各子带图像中进行块奇异值分解，求取区域特征能量值作为脉冲耦合神经网络对应神经元的链接强度。经过脉冲耦合神经网络点火处理，获取子带图像的点火映射图，通过判决选择算子，选择各子带图像中的明显特征部分生成子带融合图像；最后，应用局部化非下抽样剪切波逆变换重构图像。采用多组红外与可见光图像进行融合实验，并对融合结果进行了客观评价。实验结果表明本文提出的融合方法在主观和客观评价上均优于已有文献的一些典型融合方法，可获得更好的融合效果。
위료제승홍외여가견광도상융합정도，제출료일충국부화비하추양전절파변환여맥충우합신경망락상결합적홍외여가견광도상융합방법。수선，이용국부화비하추양전절파대원도상진행다척도、다방향분해；연후，재분해후적각자대도상중진행괴기이치분해，구취구역특정능량치작위맥충우합신경망락대응신경원적련접강도。경과맥충우합신경망락점화처리，획취자대도상적점화영사도，통과판결선택산자，선택각자대도상중적명현특정부분생성자대융합도상；최후，응용국부화비하추양전절파역변환중구도상。채용다조홍외여가견광도상진행융합실험，병대융합결과진행료객관평개。실험결과표명본문제출적융합방법재주관화객관평개상균우우이유문헌적일사전형융합방법，가획득경호적융합효과。
For enhancing fusion accuracy of infrared and visible images, an adaptive fusion algorithm of infrared and visible images based on Local Nonsubsampled Shearlet Transform (LNSST) and Pulse Coupled Neural Networks (PCNN) is proposed. First,source images are decomposed to multi-scale and multi-direction subband images by LNSST. Secondly, blocked singular value decomposition of each subband image is done to calculate the area feature energy value which is served as linking strength of each neuron in PCNN. After the processing of PCNN with the adaptive linking strength, new fire mapping images of the entire subband images are obtained, the clear objects of subband images are selected by the compare-selection operator with the fire mapping images pixel by pixel and then all of them are merged into a group of new clear subband images. Finally, fused subband images are reconstructed to image by local nonsubsampled shearlet inverse transform. Some fusion experiments on several sets of infrared and visible images are done and objective performance assessments are implemented to fusion results. The experimental results indicate that the proposed method performs better in subjective and objective assessments than a few existing typical fusion techniques in the literatures and obtains better fusion performance.