More Works from Our Group

AQIFormer: A Transformer-Based Multi-View Architecture for Cross-City Air Quality Classification

Transformer-based multi-view model that fuses front-rear traffic images and meteorological parameters for robust air quality classification and cross-city generalization.

ICVGIP 2025
IoT-based AQI Estimation using Image Processing and Learning Methods

Combines low-cost sensors and traffic images for smart air quality monitoring.

WF-IoT 2022

TRAQID: Traffic-Related Air Quality Image Dataset

Om Rajendra Kathalkar, Nitin Nilesh, Sachin Chaudhari, Anoop Namboodiri
Signal Processing and Communication Research Center (SPCRC) & Center for Visual Information Technology (CVIT)
International Institute of Information Technology, Hyderabad
Indian Conference on Computer Vision, Graphics and Image Processing (ICVGIP) 2024
Paper (ACM) Code Dataset Poster

Abstract

Air quality estimation through sensor-based methods is widely used. Nevertheless, their frequent failures and maintenance challenges constrain the scalability of air pollution monitoring efforts. Recently, it has been demonstrated that air quality estimation can be done using image-based methods. These methods offer several advantages including ease of use, scalability, and low cost. However, the accuracy of these methods hinges significantly on the diversity and magnitude of the dataset utilized.

Addressing this gap, we present TRAQID - Traffic-Related Air Quality Image Dataset, a novel dataset capturing 26,678 front and rear images of traffic alongside co-located weather parameters, multiple levels of Particulate Matters (PM) and Air Quality Index (AQI) values. Spanning over multiple seasons, with over 70 hours of data collection in the twin cities of Hyderabad and Secunderabad, India, the TRAQID offers diverse day and night imagery amid unstructured traffic conditions, encompassing six AQI categories ranging from "Good" to "Severe".

State-of-the-art air quality estimation techniques, which were trained on smaller and less-diverse datasets, showed poor results on the dataset presented in this paper. TRAQID models various uncertainty types, including seasonal changes, unstructured traffic patterns, and lighting conditions. The information from the two views (front and rear) of the traffic can be combined to improve the estimation performance in such challenging conditions. As such, the TRAQID serves as a benchmark for image-based air quality estimation tasks and AQI prediction, given its diversity and magnitude.

Dataset Overview & Collection Pipeline

TRAQID dataset overview showing diverse traffic scenes across AQI categories
Visual spectrum of TRAQID, demonstrating front and rear traffic imagery, day and night captures, and AQI diversity across six categories ranging from "Good" to "Severe" in the twin cities of Hyderabad and Secunderabad.

TRAQID data collection vehicle setup and preprocessing pipeline
Summary of TRAQID dataset creation: (a) data collection vehicle setup with dual cameras and AQ sensors, (b) street view of collection routes, (c) preprocessing steps, and (d) example data sample with features and labels.

Comparison of TRAQID with existing air quality datasets
Comparison of TRAQID with previously collected datasets, highlighting unique features including multiple views, night images, co-located data samples, sequential arrangement, weather parameters, season diversity, and dataset mobility.

GradCAM activation maps showing model attention across AQI categories
GradCAM activation map visualization of the AQC-Net model across different AQI categories. The model focuses on vehicles (especially exhaust areas), urban structures (flyovers, buildings), and atmospheric conditions. As AQI deteriorates from "Poor" to "Severe", attention shifts towards broader scene analysis indicating overall atmospheric haze.

BibTeX

@article{Kathalkar2024TRAQID,
  title     = {TRAQID - Traffic-Related Air Quality Image Dataset},
  author    = {{Om Kathalkar} and Nilesh, Nitin and Chaudhari, Sachin and Namboodiri, Anoop},
  year      = {2024},
  month     = {oct},
  publisher = {Association for Computing Machinery},
  journal   = {15th Indian Conference on Computer Vision Graphics and Image Processing, Bengaluru, India},
  doi       = {10.1145/3702250.3702260},
  url       = {https://dl.acm.org/doi/10.1145/3702250.3702260},
  dataset   = {https://india-data.org/dataset-details/4965db32-3676-427f-86f0-c8ed678dad2b}
}