Geometric Calibration of Low-Cost Flatbed Scanners for Large Scale Mapping Applications

摘要

Scanning hardcopy non-metric images is one of the most important sources in digital mapping. Low-cost scanners are still widely used in many applications as they can produce digital images of comparable precisions to those produced by expensive professional scanners. Yet, inexpensive scanners introduce geometrical distortions in the measured image coordinates that must be assessed and compensated before using their products for further analysis. In this article, several 2D-to-2D transformation models were investigated to calibrate flatbed scanners with different resolutions and sizes. We evaluated the potential of each model using two gridded-crosses plotted on high-quality transparent sheets. Control coordinates were provided through a photogram-metric analytical plotter. After scanning the sheets, least squares matching was applied to determine the precise locations of the crosses. By comparing the control coordinates and those estimated from digitized images, it was found that the mathematical model based on the projective transformation gives the best results for standardizing the geometric properties of flatbed scanners. The results show that scanning resolution of 2400 dpi achieves the requirements for large-scale mapping applications.