<p>Infrared intensity and degree of linear polarization (DoLP) image fusion enhances visual quality and detail representation in complex environments. However, existing fusion methods do not fully consider the characteristics of the DoLP images and their susceptibility to noise interference. This paper proposes an edge detail enhancement method for infrared intensity and polarization image fusion based on rolling guidance filtering (RGF) and sparse representation. Our method begins with bilateral filtering of the Stokes parameter images in the polarization vector space, effectively suppressing noise while retaining polarization signatures. Subsequently, the intensity and low-noise polarization images are decomposed into base and detail layers with the RGF. A multiscale fusion strategy combines dictionary learning and sparse representation for detail layer enhancement, coupled with visual saliency-weighted fusion of base layers to maintain the integrity of the structural information. Experimental results demonstrate that the proposed method achieves superior visual quality and enhances brightness while preserving intricate image details. Compared with existing methods, the proposed approach exhibits excellent performance in subjective visual and objective evaluation metrics.</p>