ICP-BASED SCAN ALIGNMENT

Abstract

The article explores the challenges of dental 3D scanning accuracy linked to errors that arise during the creation of three-dimensional models. These inaccuracies result from hardware limitations and constraints in data processing, which directly affect the quality of the final output — an essential factor for dental practice.

To create a solid 3D object, the authors propose a scan alignment method using the Iterative Closest Point (ICP) algorithm. This approach minimizes errors by aligning point coordinates and surface normals. The paper also introduces modifications to the ICP algorithm, including normalization and application of rotation matrices, which increase the accuracy of subscan matching.

Particular attention is paid to the mathematical framework underlying the improved algorithms, in particular to the methods of minimising the distance between corresponding scan points and accounting for the geometric parameters of surfaces. The use of weighting coefficients to reduce the influence of outlier points and noise is also discussed.

By integrating multiple scans into a single cohesive model, these improvements help reduce noise and discrepancies.

Studies have shown that the quality of the scanned points plays a key role in the algorithm's performance.

The article provides a comparative analysis of the results obtained using the classical ICP and enhanced versions of the algorithm, demonstrating a significant reduction in root mean square error. Increasing the number of cameras and capturing more detailed data significantly enhance accuracy. Experimental results have shown that the proposed techniques lead to substantial improvements in the quality of 3D models.

It is noted that the developed methods can be integrated into existing CAD/CAM systems, enabling broader application in dental practice. The importance of automating the alignment process is also emphasized, which reduces the need for manual intervention and lowers the risk of human error.

The proposed algorithm is of practical use in digital dentistry, particularly in the design of high-precision prosthetic models and other restorative procedures. Furthermore, the article identifies opportunities for further enhancements to the ICP algorithm, which will contribute to advancements in three-dimensional scanning technology in healthcare.

Key words: ICP, dentistry, alignment algorithms, 3D models, scans