基于多尺度分解和显著性区域提取的可见光红外图像融合方法

Fusion Method of Visible and Infrared Images Based on Multi-Scale Decomposition and Saliency Region Extraction

可见光红外图像融合技术对于提升成像区域的信息丰富程度具有重要意义。提出了一种基于多尺度分解和显著性区域提取的可见光红外图像融合算法。利用边缘保持的图像平滑算法,构建了多尺度图像分解框架,将图像分解为不同尺度的基础层图像和若干细节层图像,同时结合导向滤波器,在每个分解图层实施显著性区域提取。通过加权重建进行融合信息的视觉增强,得到最终的融合结果。针对不同融合算法和图像库开展了主客观评价对比实验,结果表明:所提出的算法具有较好的主客观评价结果,算法融合效果表现优异,适用性较好。

The fusion technique of visible and infrared images has important significance in enhancing the information richness of the imaging areas. A fusion algorithm of visible and infrared images based on the multi-scale decomposition and saliency region extraction is proposed. The edge-preserved image smoothing algorithm is introduced to build the framework of multi-scale image decomposition. The source image is decomposed into base layer image and several detail layer images with different decomposition scales. Meanwhile, the saliency region maps are extracted in each decomposition layer combined with the guided filter. The final fusion image is obtained by the reconstruction of each decomposition layer with different weighting factor values in order to enhance the visual effect of the fusion information. The contrast experiments of objective and subjective evaluation are developed on different fusion algorithms and databases. The experimental results illustrate that the proposed algorithm has a superior objective and subjeetive evaluation performance on the fusion results. The fusion effect of algorithm is excellent and the applicability is good.