Determination of the Quantitative Characteristics of Building Wall Defects Due to Distortion of the Projected Structured Grid

Abstract

The paper considers the computer vision system that allows to determine the type and quantitative characteristics of wall defects in buildings. The functional scheme of this system, its advantages in comparison with existing methods of defect measurements are presented. The proposed algorithm for reconstruction of the wall geometry is based on the analysis of the image of a projected structured grid. In case of defects, namely, deviation from the vertical, convexity or concavity, the structured light grid is distorted, and the geometry of the wall is restored according to the quantitative characteristics of the distortions. The quantitative characteristics of distortions of a structured light grid are determined by the deviation of the coordinates of the key points (the intersection of vertical and horizontal contrast lines) from their coordinates in the ideal case of projecting onto an absolutely smooth vertical plane. A formula for calculating the deviation from the vertical for each key point is given. The research results of the accuracy of restoring the geometric characteristics of wall defects from the type of the structured light grid are presented. The research concluded that the grid in the form of contrasting horizontal and vertical lines is preferable for restoring the quantitative parameters of the considered wall defects. The errors in determining of the geometric characteristics of defects do not exceed 7%. It is shown that even simple types of grids allow for achieving results comparable in accuracy with manual measurements, however, the measurement process is less laborious and does not depend on the “human factor”.