Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

ConPR: Ongoing Construction Site Dataset for Place Recognition

Dongjae Lee, Minwoo Jung, Ayoung Kim

TL;DR

ConPR addresses the need for place recognition benchmarks that account for terrain changes in dynamic outdoor environments, such as active construction sites. The authors introduce a multi-session dataset captured with a handheld rig containing a monocular camera, a solid-state LiDAR, an IMU, and GPS, enabling visual, LiDAR, and fused place-recognition experiments. Key contributions include the Spatial Dynamics of multiple data collections, multi-modal sensor data with ground-truth for range-based evaluation, and coverage of diverse scenarios including night-time and dynamic objects. This dataset provides a practical benchmark to evaluate and improve place-recognition algorithms under real-world, evolving conditions with sensor fusion considerations.

Abstract

Place recognition, an essential challenge in computer vision and robotics, involves identifying previously visited locations. Despite algorithmic progress, challenges related to appearance change persist, with existing datasets often focusing on seasonal and weather variations but overlooking terrain changes. Understanding terrain alterations becomes critical for effective place recognition, given the aging infrastructure and ongoing city repairs. For real-world applicability, the comprehensive evaluation of algorithms must consider spatial dynamics. To address existing limitations, we present a novel multi-session place recognition dataset acquired from an active construction site. Our dataset captures ongoing construction progress through multiple data collections, facilitating evaluation in dynamic environments. It includes camera images, LiDAR point cloud data, and IMU data, enabling visual and LiDAR-based place recognition techniques, and supporting sensor fusion. Additionally, we provide ground truth information for range-based place recognition evaluation. Our dataset aims to advance place recognition algorithms in challenging and dynamic settings. Our dataset is available at https://github.com/dongjae0107/ConPR.

ConPR: Ongoing Construction Site Dataset for Place Recognition

TL;DR

ConPR addresses the need for place recognition benchmarks that account for terrain changes in dynamic outdoor environments, such as active construction sites. The authors introduce a multi-session dataset captured with a handheld rig containing a monocular camera, a solid-state LiDAR, an IMU, and GPS, enabling visual, LiDAR, and fused place-recognition experiments. Key contributions include the Spatial Dynamics of multiple data collections, multi-modal sensor data with ground-truth for range-based evaluation, and coverage of diverse scenarios including night-time and dynamic objects. This dataset provides a practical benchmark to evaluate and improve place-recognition algorithms under real-world, evolving conditions with sensor fusion considerations.

Abstract

Place recognition, an essential challenge in computer vision and robotics, involves identifying previously visited locations. Despite algorithmic progress, challenges related to appearance change persist, with existing datasets often focusing on seasonal and weather variations but overlooking terrain changes. Understanding terrain alterations becomes critical for effective place recognition, given the aging infrastructure and ongoing city repairs. For real-world applicability, the comprehensive evaluation of algorithms must consider spatial dynamics. To address existing limitations, we present a novel multi-session place recognition dataset acquired from an active construction site. Our dataset captures ongoing construction progress through multiple data collections, facilitating evaluation in dynamic environments. It includes camera images, LiDAR point cloud data, and IMU data, enabling visual and LiDAR-based place recognition techniques, and supporting sensor fusion. Additionally, we provide ground truth information for range-based place recognition evaluation. Our dataset aims to advance place recognition algorithms in challenging and dynamic settings. Our dataset is available at https://github.com/dongjae0107/ConPR.
Paper Structure (6 sections, 2 figures)

This paper contains 6 sections, 2 figures.

Table of Contents

  1. introduction & related works
  2. System Overview
  3. Sensor Configuration
  4. Sensor Calibration
  5. Ongoing Construction Site Dataset
  6. conclusion

Figures (2)

  • Figure 1: These sample RGB images and 3D point cloud maps are extracted from our ongoing construction site dataset. The dataset captures the development of buildings and bridges in the active construction site. In addition to showcasing the construction progress, this dataset presents various challenges, including the presence of dynamic objects like vehicles, pedestrians, and construction equipment, along with variations in lighting conditions and viewpoints.
  • Figure 2: Handheld system incorporating a monocular camera and solid-state LiDAR.