Against the backdrop of the global oil and gas industry's strategic shift towards deepwater exploration, marine acquisition technologies are continually evolving. Ocean Bottom Node (OBN) acquisition has emerged as a key technique, offering advantages over conventional towed-streamer methods such as flexible deployment, rich low-frequency content, and full-azimuth coverage, making it essential for imaging deep and complex reservoirs. However, the benefits of multi-component OBN acquisition also introduce significant processing challenges, notably the shear-wave noise unique to the vertical geophone (Z) component and the ghost wave interference caused by sea-surface reflections in both the hydrophone and geophone data. This paper presents targeted techniques for jointly attenuating shear noise and ghost waves: firstly, a 3D sparse transform and complex wavelet domain joint shear-wave noise model is established to achieve precise noise attenuation; subsequently, based on the high-fidelity characteristics of the hydrophone and geophone data, a novel refracted wave matching algorithm is innovatively employed to perform dual-sensor combination for effective deghosting. The proposed methods were applied to real data processing from a block in the western South China Sea, delivering effective noise suppression and significantly improving the imaging quality of weak reflectors in the mid-to-deep layers.