Related Experiment Video
Updated: Oct 1, 2025
Author Spotlight: UAV Remote Sensing for Efficient Invasive Plant Biomass Estimation
Published on: February 9, 2024
A PM2.5 concentration estimation method based on multi-feature combination of image patches
Xiaochu Wang1, Meizhen Wang1, Xuejun Liu1
1School of Geography, Nanjing Normal University, Nanjing, 210023, China; Key Laboratory of Virtual Geographic Environment, Nanjing Normal University, Ministry of Education, Nanjing, 210023, China; Jiangsu Center for Collaborative Innovation in Geographical Information Resource Development and Application, Nanjing Normal University, Nanjing, 210023, China.
This study introduces an image-based method for estimating fine particulate matter (PM2.5) concentration. The approach enhances accuracy by incorporating environmental factors and a novel patchwise strategy, proving effective across diverse datasets.
Area of Science:
- Environmental Science
- Remote Sensing
- Data Science
Background:
- Accurate monitoring of fine particulate matter (PM2.5) is crucial for environmental protection and public health.
- Existing PM2.5 estimation methods face challenges in efficiency, accuracy, and resolution.
- Developing novel approaches for real-time and high-resolution PM2.5 monitoring is a significant research area.
Purpose of the Study:
- To propose and validate an efficient, accurate, and high-resolution image-based method for estimating PM2.5 concentrations.
- To investigate the impact of incorporating environmental influence factors on PM2.5 estimation accuracy.
- To enhance the predictive performance of PM2.5 estimation models using an improved patchwise strategy.
Main Methods:
- Developed an image-based method for PM2.5 concentration estimation.
- Integrated image features with environmental factors (e.g., relative humidity, time of year) for improved inference.
- Employed an improved patchwise strategy for regression and prediction processes.
- Validated the method using the Shanghai scene dataset and two additional datasets from different times and locations.
Main Results:
- The proposed method achieved high estimation accuracy, with R² of 0.88 and RMSE of 10.42 μg·m⁻³ on the Shanghai dataset.
- Incorporating influence factors like relative humidity and photographing month significantly improved estimation accuracy.
- The improved patchwise strategy demonstrated a substantial enhancement in predictive performance.
- Cross-validation on diverse datasets confirmed the method's effectiveness and broad applicability.
Conclusions:
- The image-based approach offers an efficient and accurate solution for PM2.5 monitoring.
- Environmental factors and advanced strategies like patchwise processing are key to improving PM2.5 estimation.
- The proposed method shows promise for real-world applications in air quality assessment and public health initiatives.
