Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

SID: Stereo Image Dataset for Autonomous Driving in Adverse Conditions

Zaid A. El-Shair, Abdalmalek Abu-raddaha, Aaron Cofield, Hisham Alawneh, Mohamed Aladem, Yazan Hamzeh, Samir A. Rawashdeh

TL;DR

The paper introduces SID, a large-scale stereo-image dataset designed to advance autonomous driving perception under adverse weather and lighting. Collected with a ZED stereo camera at 20 Hz across 27 sequences, SID provides over 178k stereo image pairs with rich sequence-level annotations (weather, time of day, location, road conditions) and includes challenging lens soiling scenarios. It fills a gap left by prior datasets that lack stereo-depth data and comprehensive adverse-condition coverage, enabling robust depth perception, weather-aware analysis, and cross-condition benchmarking with a 74/26 train/test split. Publicly available data and detailed documentation support research into robust perception, depth estimation, and autonomous navigation in real-world, variable environments.

Abstract

Robust perception is critical for autonomous driving, especially under adverse weather and lighting conditions that commonly occur in real-world environments. In this paper, we introduce the Stereo Image Dataset (SID), a large-scale stereo-image dataset that captures a wide spectrum of challenging real-world environmental scenarios. Recorded at a rate of 20 Hz using a ZED stereo camera mounted on a vehicle, SID consists of 27 sequences totaling over 178k stereo image pairs that showcase conditions from clear skies to heavy snow, captured during the day, dusk, and night. The dataset includes detailed sequence-level annotations for weather conditions, time of day, location, and road conditions, along with instances of camera lens soiling, offering a realistic representation of the challenges in autonomous navigation. Our work aims to address a notable gap in research for autonomous driving systems by presenting high-fidelity stereo images essential for the development and testing of advanced perception algorithms. These algorithms support consistent and reliable operation across variable weather and lighting conditions, even when handling challenging situations like lens soiling. SID is publicly available at: https://doi.org/10.7302/esz6-nv83.

SID: Stereo Image Dataset for Autonomous Driving in Adverse Conditions

TL;DR

The paper introduces SID, a large-scale stereo-image dataset designed to advance autonomous driving perception under adverse weather and lighting. Collected with a ZED stereo camera at 20 Hz across 27 sequences, SID provides over 178k stereo image pairs with rich sequence-level annotations (weather, time of day, location, road conditions) and includes challenging lens soiling scenarios. It fills a gap left by prior datasets that lack stereo-depth data and comprehensive adverse-condition coverage, enabling robust depth perception, weather-aware analysis, and cross-condition benchmarking with a 74/26 train/test split. Publicly available data and detailed documentation support research into robust perception, depth estimation, and autonomous navigation in real-world, variable environments.

Abstract

Robust perception is critical for autonomous driving, especially under adverse weather and lighting conditions that commonly occur in real-world environments. In this paper, we introduce the Stereo Image Dataset (SID), a large-scale stereo-image dataset that captures a wide spectrum of challenging real-world environmental scenarios. Recorded at a rate of 20 Hz using a ZED stereo camera mounted on a vehicle, SID consists of 27 sequences totaling over 178k stereo image pairs that showcase conditions from clear skies to heavy snow, captured during the day, dusk, and night. The dataset includes detailed sequence-level annotations for weather conditions, time of day, location, and road conditions, along with instances of camera lens soiling, offering a realistic representation of the challenges in autonomous navigation. Our work aims to address a notable gap in research for autonomous driving systems by presenting high-fidelity stereo images essential for the development and testing of advanced perception algorithms. These algorithms support consistent and reliable operation across variable weather and lighting conditions, even when handling challenging situations like lens soiling. SID is publicly available at: https://doi.org/10.7302/esz6-nv83.
Paper Structure (17 sections, 2 equations, 7 figures, 1 table)

This paper contains 17 sections, 2 equations, 7 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. Data Collection
  4. Data Collection Methodology and Hardware
  5. Environmental Conditions and Locations
  6. Dataset Description and Analysis
  7. Dataset Description
  8. Structure and Decomposition
  9. Data Annotation
  10. Statistical Breakdown and Data Insights
  11. Applications and Use-Cases
  12. Weather Classification and Understanding
  13. Robust Stereo Vision Algorithms
  14. Day/Night Image Enhancement
  15. Autonomous Vehicle Navigation and Control Systems
  16. ...and 2 more sections

Figures (7)

  • Figure 1: Stereo image sample showing the left and right images captured by the ZED camera at 720p resolution.
  • Figure 2: Multiple weather conditions included in our dataset such as (a) Clear, (b) Overcast, (c) Rain, and (d) Snow.
  • Figure 3: Various road conditions present in our dataset such as (a) Dry, (b) Wet, (c) Snow and Wet, and (d) Snowy.
  • Figure 4: Locations available in our dataset including (a) Campus (CW), (b) Campus (CCW), (c) Residential, and (d) City.
  • Figure 5: SID sequence distribution across various metrics.
  • ...and 2 more figures