In order to enhance underwater image quality in turbid water environments, a polarization image restoration method for concrete structures in turbid water was proposed based on laboratory underwater imaging experiments. The polarization characteristic difference between the concrete’s structural information light and backscattered light was analyzed, and the Stokes vector was used to obtain the polarization parameter image of the underwater image. The contrast-limiting adaptive histogram equalization method was then employed to enhance image contrast. Finally, the underwater polarization imaging model was utilized to achieve image restoration. Comparative experiments against ten image processing methods demonstrate that in highly turbid sediment-laden water environments, the proposed method achieves improvement rates of 252%, 561%, 241%, and 93% for the four evaluation metrics: average gradient, entropy enhancement, underwater color image quality assessment, and underwater image quality measurement, respectively. This enables the acquisition of underwater images of apparent defects in concrete structures with high contrast and clear details, making it suitable for restoring underwater images of apparent defects in concrete structures.