Super-Resolution Reconstruction of Color Mural Image Based on Sparse Representation of Quaternary Multiple Features
Color mural images usually contain complex textures and rich color information.The traditional super-resolution im-age reconstruction method based on channel separation only processes the luminance information of color murals images,ignoring the numerous color detail information it contains and the correlation information between channels,which easily results in blurred edges and color artifacts in the reconstructed images.To address these issues,this paper combines the information of three channels of Lαβ color space with local variance and applies it to the quaternion model to characterize the details,luminance,and color infor-mation of color mural images.Based on this,a joint multi-feature dictionary model based on detail-brightness-color information is constructed to enhance the detail information of the reconstruction results while effectively using the color information of each chan-nel.The experimental results show that the method performs better in both subjective and objective indexes in the reconstruction of color mural images.
super-resolutionsparse representationquaternionLαβ color space
万方数据
维普
