VIF–TOPSIS coupling algorithm for image quality assessment in smart construction site management

Abstract

Real-time monitoring is critical for smart construction management, yet environmental complexities degrade surveillance video quality. Traditional visual information fidelity (VIF) algorithms depend on reference images, which limits their use in no-reference scenarios such as autonomous systems and industrial inspection. Grounded in information theory, this paper proposes an algorithm that integrates VIF with the technique for order preference by similarity to ideal solution (TOPSIS), obtaining the coupling algorithm VIF–TOPSIS. Wavelet transforms extract the features; TOPSIS selects the optimal features; and the inverse wavelet transform reconstructs the images. An “ideal solution” image replaces the reference dependencies, enabling no-reference quality assessment. Evaluated on multiframe samples, the algorithm significantly improved peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), and gradient magnitude similarity deviation (GMSD) compared to traditional methods. Enhanced contrast and brightness further validated its efficacy in dynamic environments. This framework advances real-time video processing, offering robust technical support for smart construction management.