<p>In printing-oriented document image processing, color serves as a pivotal prior for faithful texture restoration. The distinctive chromatic signatures of color annotations, seals, and charts not only encode semantic cues but also provide multi-dimensional constraints for high-frequency detail reconstruction. Nevertheless, existing document-image super-resolution methods remain confined to global optimization within a single color space, thereby entangling color and texture representations and capping reconstruction fidelity. To overcome this limitation, we propose a texture and chrominance dual-guided super-resolution framework for complex documents. A parallel reconstruction mechanism is introduced that concurrently performs texture recovery in the RGB domain and chrominance refinement in the CIELAB space, circumventing the coupling artifacts inherent to single-color-space approaches. Specifically, in RGB, a texture-priority reconstruction network with hierarchical residual blocks is designed to restore high-frequency structural details under explicit gradient fidelity constraints. In CIELAB, a multi-level skip-connection fusion module augmented by a hybrid channel spatial attention mechanism adaptively re-weights feature maps according to channel-wise chrominance sensitivity and local spatial gradients, enabling dynamic color restoration. Furthermore, a coupled texture-chrominance loss, regularized by cross-domain feature fusion penalties, enforces mutual guidance between the two branches, amplifying chrominance induced texture enhancement. Experimental results on both public benchmarks and a newly curated printing-scenario dataset demonstrate state-of-the-art performance, substantiating the efficacy of the proposed approach.</p>