Dataset: Image-Derived Chromatic and Photometric Descriptors as Proxies for Urban Air Quality Parameters: A Smartphone-Based Remote Sensing Approach

Description

Atmospheric raw data from RMCAB (from the Spanish acronym for Air Quality Monitoring Network of Bogotá) and JPG landscape images of urban zone in Bogotá - Colombia.

Steps to reproduce

Initially, the images were classified according to atmospheric conditions using a Normalised Blue Fraction Index (NBFI), NBFI=(〖#P〗_Blue/〖#P〗_Total )/(Max (〖#P〗_Blue/〖#P〗_Total ) ); where 〖#P〗_Blue denotes the number of pixels with the RGB colour value (0.405418, 0.598949, 0.903125), corresponding to sky blue within the images; and 〖#P〗_Total denotes the total number of pixels in each image, retaining only clear-sky and partly cloudy observations for further analysis. The selected images were then segmented into three Regions of Interest (RoIs) representing the sky, mountain background and urban structures. The RGB channels were subsequently linearised to remove the gamma correction inherent to the sRGB colour space. Chromatic and photometric descriptors were then extracted from the RGB, HSV, CIE XYZ and CIELAB colour spaces. Median values were calculated for each descriptor to reduce the influence of outliers, and hourly medians were subsequently computed to match the temporal resolution of the atmospheric monitoring stations. The image-derived descriptors were subsequently integrated with the atmospheric observations through structured data matrices constructed for each Region of Interest and atmospheric scenario. These integrated datasets constituted the input required for the subsequent statistical analyses, enabling direct comparison between image-derived information and instrumental air quality measurements through Spearman's rank correlation coefficient with a significance level of α = 0.05. Finally, the magnitude of the correlations was interpreted according to Cohen's classification, allowing the identification of the chromatic and photometric descriptors exhibiting the strongest and most consistent associations with instrumental air quality measurements.

Institutions

• Fundación Universitaria Los Libertadores

  Bogota D.C., Bogotá

Categories

Licence