Output details
14 - Civil and Construction Engineering
University College London
A comparison of dense matching algorithms for scaled surface reconstruction using stereo camera rigs
Dense image matching can be used to create point clouds comparable in density to laser scanners, but at much lower sensor costs. This paper investigates the geometric accuracy of the resulting point clouds in a metrology context using industrial guidelines. The paper for the first time compares four recently developed matching algorithms w.r.t. their accuracy for industry application. The findings are important e.g. for designing inspection systems for production industry. They help in balancing sensor cost, algorithmic performance and accuracy. The paper was published in the highest ranked journal for photogrammetry earlier this year and has already attracted citations (Scopus).