Integration of Remote Sensing and Multivariate Linear Regression Analysis to Evaluate the Air Pollution Index in Binh Duong Province, Vietnam
摘要
This study was conducted to (1) build a multivariate linear regression model for the API (Air Pollution Index) from data from Landsat 8 images and periodic environmental monitoring and (2) establish maps of air pollution zoning in Binh Duong province based on API. Research results show that 5/12 input variables, including Band 6, Band 7, NDVI, VI, and Band 1, are statistically significant with their p value < 0.05. The best-fitting regression function has high values of R (0.897) and Radj (0.895) and low Root Mean Square Error (RMSE = 7.3). The air pollution map shows that 84.5% of Binh Duong province has clean quality, 15.4% of the area has average quality, and only 0.1% of the areas are slightly polluted, and there are no severely or seriously polluted areas in Binh Duong province. This is valuable information for air quality management based on remote sensing data in Binh Duong province.