Optimizing overlap percentage for enhanced accuracy and efficiency in oblique photogrammetry building 3D modeling
Size Wu, Feng Lin, Xiaorui Zhang, Chaoen Yin, Quan Lei, Bo Tian
Abstract
This paper investigates the impact of image overlap in oblique photogrammetry on the size of the dataset and the accuracy of the resulting 3D model, which is crucial for optimizing efficiency and cost-effectiveness in large-scale 3D reconstruction projects. By introducing a regression mathematical model, we establish a relationship between the overlap percentage, dataset size, and model accuracy. Through validation using actual flight data, we demonstrate that an overlap percentage exceeding 70 % heading and 60 % lateral is necessary to meet practical requirements. Our findings reveal a nonlinear improvement in model accuracy as overlap increases, with notable gains at 80 % heading and 75 % lateral overlap, offering a 40 % accuracy enhancement at a 30 % data volume increase compared to 70 % heading and 60 % lateral overlap. However, excessive overlap (90 % heading and 80 % lateral) leads to marginal accuracy gains while significantly increasing data volume by 300 %. Balancing these factors, we recommend an optimal overlap range of 75 %-85 % heading and 70 %-80 % lateral for efficient and high-quality 3D modeling using oblique photogrammetry.