COMPASS: COmpact Multi-channel Prior-map And Scene Signature for Floor-Plan-Based Visual Localization

2026-04-28

AI Summary

This work addresses the limitation of existing visual localization methods that predominantly rely on geometric structures while neglecting semantic information in building floor plans. To bridge this gap, the authors propose a cross-modal localization approach that integrates both geometric and semantic priors, introducing semantic elements such as windows for the first time into the matching process between floor plans and fisheye images. They design a five-channel radial descriptor—incorporating normalized distance, structural hit type, distance gradient, and other cues—and combine it with a fisheye camera model, line segment detection, vertical edge clustering, and brightness validation to construct a compact scene representation. Single-frame experiments on the Hilti-Trimble SLAM Challenge 2026 dataset demonstrate high consistency between visual observations and floor plans using the proposed descriptor, confirming the method’s effectiveness and feasibility.

Abstract

Architectural floor plans are widely available priors which contain not only geometry but also the semantic information of the environment, yet existing localization methods largely ignore this semantic information. To address this, we present COMPASS, an algorithm that exploits both geometric and semantic priors from floor plans to estimate the pose of a robot equipped with dual fisheye cameras. Inspired by scan context descriptor from LiDAR-based place recognition, we design a multi-channel radial descriptor that encodes the geometric layout surrounding a position. From the floor plan, rays are cast in 360 azimuth bins and the results are encoded into five channels: normalized range, structural hit type (wall, window, or opening), range gradient, inverse range, and local range variance. From the image side, the same descriptor structure is populated by detecting structural elements in the fisheye imagery. As a first step toward full cross-modal matching, we present a window detection algorithm for fisheye images that uses a line segment detector to identify window frames via vertical edge clustering and brightness verification. Detected windows are projected to azimuthal bearings through the fisheye camera model, producing the hit-type channel of the visual descriptor. As a proof of concept, we generate both descriptors at a single known pose from the Hilti-Trimble SLAM Challenge 2026 dataset and demonstrate that the wall-window pattern extracted from the first frame of each camera closely matches the floor plan descriptor, validating the feasibility of cross-modal structural matching.

Problem

• floor-plan-based visual localization
• semantic priors
• geometric layout
• cross-modal matching
• fisheye imagery

Innovation

• floor-plan-based localization
• multi-channel radial descriptor
• cross-modal matching
• fisheye window detection
• structural scene signature