Min-kyeong Seo, Jong-taek Oh, Journal of Internet Computing and Services, Vol. 27, No. 1, pp. 113-120, Feb. 2026
10.7472/jksii.2026.27.1.113, Full Text:  HTML
Keywords: QR code, information decoding, vertex coordinate, camera position, image processing

Abstract

Research has been conducted on using QR code images for recognizing mobile object positions. However, as the distance between the camera and the QR code increases, the information contained in the QR code becomes undecodable, and the error in estimating the coordinates of the four corners of the QR code increases, leading to larger errors in camera position estimation. This paper presents research aimed at increasing the decoding rate of information contained in captured QR codes and improving the accuracy of the pixel positions of the position detection patterns within the QR code, ultimately enhancing the accuracy of camera position estimation. By combining various image processing algorithms, dozens of image pre-processing pipelines were created. The original raw images were processed through these pipelines, and then conventional QR code recognition programs were used to decode the information and estimate the corner coordinates. By developing and adjusting various image pre-processing pipelines and their parameters, a significant 19.2% improvement in the success rate of QR code information decoding was achieved compared to using the raw images. The accuracy of corner coordinate estimation also improved, resulting in an approximately 90% improvement in the final camera position estimation accuracy. The results of this study can be applied to all applications that require precise indoor and outdoor position estimation using QR code images attached to walls or mobile objects, and are characterized by a very high cost-effectiveness ratio.

Statistics
Cite this article