Imaging air quality evaluation using definition metrics and detrended fluctuation analysis

This paper presents a novel image quality evaluation method which addresses the air quality evaluation issue specifically. The fluctuation of the imaging definition can represent the change laws of the air quality indirectly. So, first three types of non-reference Image Quality (IQ) metric, i.e., the contrast, the blur and the noise factors, are computed for each frame of an image sequence to describe the imaging definition. Then we use these discrete IQ data to form their time series data respectively. After that, we use the Detrend Fluctuation Analysis (DFA) technique to mine the laws of fluctuation of those series. Finally, the scaling exponents of the DFA can be used to represent and classify the intensity of fluctuation of the different air quality conditions. The sequences of different atmosphere conditions are used to test the validity our model in this paper.