Air quality measurement using remote sensing and digital images processing techniques / Lim, H. S. … [et al.]

A conventional digital camera Kodak DC 290 was used as a sensor to capture digital images of our reference targets. The digital images were separated into three bands namely red, green and blue bands for multispectral algorithm calibration. In-situ measurements of corresponding air pollution parameters were carried out at the ASMA air pollution station in Universiti Sains Malaysia campus, Penang. The selected parameter was particulate matter less than 10 micron (PMlO). We put up four different colour paper (Red, Green, Blue and Black) on the wall of a building as reference surfaces. Digital images were captured at close and far distance from the colour papers/wall of the building. These in-situ measurements were then used as the dependent variables in deriving the air quality information using the digital camera data. In fact, air quality parameters can be measured using ground instrument such as air sampler. But these instruments are quite expensive, and a limited number of the air pollutant stations are available in each area. Most countries have established a network for measuring the air quality levels. Due to the high cost and limited number of the air pollutant stations in the each area, they cannot provide a good spatial distribution of the air pollutant reading over a city. The objective of this study was to test the potential of using remote sensing and digital image processing techniques for the air quality measurements. The digital numbers of three bands were converted into irradiance. and then reflectance. The relationship between the reflectance and the corresponding air quality data was determined using regression analysis. A new algorithm was developed for detecting air pollution from the digital camera images chosen based on the highest correlation coefficient, R and lowest root mean square error, RMS for PM10. The algorithm used was based on the apparent reflectance values detected at near and far distances from the refence surface, and these in turn can be related to the concentration of the air pollutants. The coefficients of the calibrated algorithm were determined and used in estimating the air pollution level. The newly developed algorithm produced a high degree of accuracy with the correlation coefficient (R) of 0.9001 and the root-mean-square error (RMS) of 5.1008 µg/m³ for PM10. Comparison was made between the air pollution parameter and the results using different colour paper or wall of a building as a reference surface. We found that the red colour paper produced the best result when it was used as a reference surface in this study. The pollutant's concentration can be estimated accurately at a relatively much cheaper cost compared with other techniques.