BIM-Constrained Optimization for Accurate Localization and Deviation Correction in Construction Monitoring

TL;DR

This work tackles drift between BIM geometry and real-time SLAM in AR-based construction monitoring, which degrades alignment on textureless and dynamic sites. It introduces a BIM-aware drift correction framework that aligns as-built planes with as-planned BIM planes and estimates the transformation between BIM and SLAM origin frames $^{B}T_S$ by solving a least-squares optimization, e.g., $\arg\min_{R,t} \sum_{i=1}^n \| R s_i + t - b_i \|^2 + \| R n_s - n_b \|^2$. The key contributions include BIM-based plane extraction and matching using Mahalanobis distance filtering with threshold $\tau$, a Gauss-Newton optimization with Dense SVD to update the transformation, and real-world validation showing substantial drift reductions (52.24% angular, 60.8% distance) across construction sites. This approach enables robust, long-term AR visualization in construction environments and points toward lightweight, mobile-friendly localization that leverages architectural BIM as a prior.

Abstract

Augmented reality (AR) applications for construction monitoring rely on real-time environmental tracking to visualize architectural elements. However, construction sites present significant challenges for traditional tracking methods due to featureless surfaces, dynamic changes, and drift accumulation, leading to misalignment between digital models and the physical world. This paper proposes a BIM-aware drift correction method to address these challenges. Instead of relying solely on SLAM-based localization, we align ``as-built" detected planes from the real-world environment with ``as-planned" architectural planes in BIM. Our method performs robust plane matching and computes a transformation (TF) between SLAM (S) and BIM (B) origin frames using optimization techniques, minimizing drift over time. By incorporating BIM as prior structural knowledge, we can achieve improved long-term localization and enhanced AR visualization accuracy in noisy construction environments. The method is evaluated through real-world experiments, showing significant reductions in drift-induced errors and optimized alignment consistency. On average, our system achieves a reduction of 52.24% in angular deviations and a reduction of 60.8% in the distance error of the matched walls compared to the initial manual alignment by the user.

Figures (5)

• Figure 1: SLAM-Based Plane Detection and BIM Alignment for AR Visualization. From the top view, different camera poses are shown, and in the right side, real-world camera images are displayed with detected planes.
• Figure 2: Overview of the system. Our approach extracts planar structures from SLAM and BIM models, applies robust plane matching, and computes a transformation (TF) using a Least Squares-based optimization method to minimize drift.
• Figure 3: Defined reference frames for this work.
• Figure 4: Comparison of rotation and translation errors across two environments using three transformation estimation variants.
• Figure 5: Comparison of camera pose trajectories. The left figure shows significant drift with the baseline method, colliding the walls in BIM, while the right one demonstrates the improved alignment with the Local Transformation approach.