Uncovering local aggregated air quality index with smartphone captured images leveraging efficient deep convolutional neural network

TL;DR

This work tackles estimating location-specific PM2.5 and equivalent daily AQI from smartphone images in Dhaka using a custom deep convolutional neural network (PPPC). It builds a public dataset of 1,818 images with NowCast-based PM2.5 labels and demonstrates that a lightweight DCNN with about $4.85\times 10^6$ parameters can outperform larger architectures such as ViT and INN for this task, leveraging image cues like transmission, sky hue, gradient, entropy, RMS contrast, and humidity. The paper provides a clear preprocessing pipeline (sky-focused cropping to 120×200, normalization) and shows moderate predictive power ($R^2 \approx 0.39$, $MAE \approx 29.7$, $RMSE \approx 42.2$), indicating practical potential for scalable, low-cost AQI estimation in resource-constrained settings. Publicly releasing code and data, the authors outline plans for mobile deployment and future multimodal improvements (e.g., temperature, humidity, wind) to enhance generalization beyond Dhaka.

Abstract

The prevalence and mobility of smartphones make these a widely used tool for environmental health research. However, their potential for determining aggregated air quality index (AQI) based on PM2.5 concentration in specific locations remains largely unexplored in the existing literature. In this paper, we thoroughly examine the challenges associated with predicting location-specific PM2.5 concentration using images taken with smartphone cameras. The focus of our study is on Dhaka, the capital of Bangladesh, due to its significant air pollution levels and the large population exposed to it. Our research involves the development of a Deep Convolutional Neural Network (DCNN), which we train using over a thousand outdoor images taken and annotated. These photos are captured at various locations in Dhaka, and their labels are based on PM2.5 concentration data obtained from the local US consulate, calculated using the NowCast algorithm. Through supervised learning, our model establishes a correlation index during training, enhancing its ability to function as a Picture-based Predictor of PM2.5 Concentration (PPPC). This enables the algorithm to calculate an equivalent daily averaged AQI index from a smartphone image. Unlike, popular overly parameterized models, our model shows resource efficiency since it uses fewer parameters. Furthermore, test results indicate that our model outperforms popular models like ViT and INN, as well as popular CNN-based models such as VGG19, ResNet50, and MobileNetV2, in predicting location-specific PM2.5 concentration. Our dataset is the first publicly available collection that includes atmospheric images and corresponding PM2.5 measurements from Dhaka. Our codes and dataset are available at https://github.com/lepotatoguy/aqi.

Figures (7)

• Figure 1: The radiance captured by the smartphone camera encompasses the combined light transmitted from the subject and the air-light from the sun, which undergoes dispersion by air, moisture, and atmospheric particles Liu2016-pc
• Figure 2: Sample image from the dataset (the image is captured by one of the authors using a smartphone)
• Figure 3: Histogram of PM2.5 of our dataset (the figure is created using Matplotlib library of Python version 3.6.13)
• Figure 4: Illustration of the proposed deep CNN architecture
• Figure 5: Loss curves of the proposed model (the figures are created using Matplotlib library of Python version 3.6.13)