基于改进的最大熵算法与滚动引导滤波的图像融合算法

Image Fusion Based on Improved Maximum Entropy Segmentation Algorithm and Rolling Guidance Filter

随着应用环境的改变,大多数红外与可见光图像融合算法的工程化部署能力普遍较差,存在目标提取不充分、细节丢失、算法复杂、效率低和适用性差等问题。针对上述问题,提出了一种基于改进的最大熵算法与滚动引导滤波的红外与可见光图像融合算法。首先,使用改进的最大熵算法对红外目标进行提取,并利用滚动引导滤波的尺度感知和边缘保持特性将可见光图像和红外图像分解为基础层和细节层;然后,提取出的红外目标和可见光基础层图像通过基础层融合规则得到基础层融合图像;最后,基础层融合图像通过细节层融合规则得到最终的融合图像。实验结果表明,所提算法的融合图像目标明确、纹理清晰、细节信息丰富,且算法简单高效、适用性强。相比其他4种对比算法,所提算法在主、客观评价上均有优势,具有一定的工程化部署能力。

With the change of application environment,the engineering deployment capability of numerous fusion algorithms for infrared and visible images is generally poor.These algorithms exhibit various shortcoming,including insufficient target extraction,loss of details,algorithmic complexity,low efficiency,and limited applicability.To address the aforementioned challenges,a fusion method for infrared and visible images is proposed by leveraging an improved maximum entropy algorithm and a rolling guided filter.First,infrared targets are extracted through using improved maximum entropy algorithm,and the visible image and infrared image are decomposed into basic layer and detail layer through using the scale-awareness and edge-preservation characteristics of rolling guided filter.Then,the base layer fusion image is generated from the extracted infrared targets and the base layer image of visible light using the base layer fusion rules.Finally,the final fusion image is derived from the base layer fusion image using the detail layer fusion rules.Experimental results evidently demonstrate that the proposed algorithm yields fused images with well-defined targets,distinct textural details,and abundant information.Moreover,the proposed algorithm stands out for its simplicity,efficiency,and broad applicability.Compared with the other four algorithms,the proposed algorithm demonstrates superior performance in both subjective and objective evaluations,andhas certain engineering gdeployment capability.