A satellite-based, near real-time, street-level resolution air pollutants monitoring system using machine learning for personalised skin health applications

AbstractSkin exposome encapsulates all internal and environmental exposures that affect skin health. Of these, photo-pollution refers to the combined effect on human skin of the simultaneous exposure to solar radiation (especially UV) and air pollution. Providing personalised photo-pollution exposure warnings and dose monitoring to an individual through a smartphone app could help in reducing skin ageing and degradation as well as in managing skin conditions (for example Atopic Dermatitis). However, accurate monitoring is challenging without a potentially expensive or cumbersome sensor device. In this work we present an innovative satellite-based air pollutant monitoring software service, ExpoPol, developed by siHealth Ltd. ExpoPol synthesises several inputs including live satellite imagery in real-time into an artificial intelligence (AI) model to provide assessment of the exposure of a smartphone user to relevant air pollutants, such as nitrogen oxides (NOx), poly-aromatic hydrocarbons (PAH) and ozone (O3). When combined with siHealth ’s patented technology HappySun® for solar radiation monitoring, ExpoPol can effectively provide a sensor-less personal skin photo-pollution dosimetry. By downscaling satellite data using local geographic data, ExpoPol is capable of monitoring pollutants with street-level resolution and global co verage in near real-time. We evaluate the accuracy of ExpoPol against ground-station monitoring data for three pollutants across three continental...
Source: Air Quality, Atmosphere and Health - Category: Environmental Health Source Type: research