A unified normalization method for homography estimation using combined point and line correspondences. (11th November 2021)

• Record Type: Journal Article
• Authors: Lu, Linjun; Dai, Fei
• Is Part Of: Computer-aided civil and infrastructure engineering. Volume 37:Number 8(2022)
• Page Start: 1010
• Page End: 1026
• Publication Date: 2021-11-11
• Subjects: Civil engineering -- Data processing -- Periodicals; Computer-aided engineering -- Periodicals; 624.0285
• DOI: 10.1111/mice.12788
• Languages: English
• ISSNs: 1093-9687

Abstract

Abstract: Data normalization is an essential and imperative step when using the direct linear transform method for homography estimation. However, the existing data normalization methods only rely on either point coordinates or line coefficients in computing homography, which is inapplicable to many civil infrastructure scenes, in which both points and lines exist but are scarce. It is plausible to maximize the accuracy level achievable by fully utilizing all points and lines available in these scenes. To fit this purpose, this study proposed a unified data normalization method for homography estimation by processing combined point and line correspondences. In this method, an analytical procedure was created to transform line coefficients for normalization under similarity transformation. By this procedure, formulas for defining parameters of the similarity transformation were developed and equations for pre‐ and post‐processing of data were established, leading to a pipeline that allows for solving different combinations of point and line correspondences simultaneously. Both simulation experiments and field tests were conducted to evaluate the performance of the method. The results showed that the method can effectively exploit all available point and line correspondences and significantly improve the accuracy for homography estimation under different measurement conditions.