Tensor-based Metal Mesh Image Denoising for Lighting Transmission Analysis

This research proposes the Tensor-based metal mesh windows denoising in the wavelet domain (TMDW) with the optical inspection to analyze the lighting transmission through the metal mesh window, which involves two main parts. The first part is an image acquisition system, which includes a high-resolution industrial camera, a lighting source that simulates the movement of the sun, a mechanical system for holding and moving metal mesh window samples, and a computer. The second part is a software system that analyzes the lighting through metal mesh windows. First, the metal mesh window image is denoised using the TMDW algorithm. Next, the denoised metal mesh image's binary image is obtained to calculate the lighting transmission through a metal mesh window. Finally, lighting transmission through the metal mesh window is calculated on the entire window and each mesh cell. The experimental results on variable samples reveal that the lighting transmission analysis task for the metal mesh window can be obtained with high accuracy and saving time processing. Therefore, the proposed system is used effectively to analyze lighting transmission in metal mesh production to ensure product quality before mass production. Besides the system is also deployed to visually inspect products during production. Not only that, but our proposed system also helps architects have a scientific basis for developing optimal lighting building designs.