Aerosol optical depth (AOD) data derived from satellite products have been widely used to estimate fine particulate matter (PM2.5) concentrations. However, existing approaches to estimate PM2.5 concentrations are invariably limited by the availability of AOD data, which can be missing over large areas due to satellite measurements being obstructed by, for example, clouds, snow cover or high concentrations of air pollution. In this study, we addressed this shortcoming by developing a novel method for determining PM2.5 concentrations with high spatial coverage by integrating AOD-based estimations and smartphone photograph-based estimations. We first developed a multiple-input fuzzy neural network (MIFNN) model to measure PM2.5 concentrations from smartphone photographs. We then designed an ensemble learning model (AutoELM) to determine PM2.5 concentrations based on the Collection-6 Multi-Angle Implementation of Atmospheric Correction AOD product. The R2 values of the MIFNN model and AutoELM model are 0.85 and 0.80, respectively, which are superior to those of other state-of-the-art models. Subsequently, we used crowdsourced smartphone photographs obtained from social media to validate the transferability of the MIFNN model, which we then applied to generate smartphone photograph-based estimates of PM2.5 concentrations. These estimates were fused with AOD-based estimates to generate a new PM2.5 distribution product with broader coverage than existing products, equating to an average increase of 12% in map coverage of PM2.5 concentrations, which grows to an impressive 25% increase in map coverage in densely populated areas. Our findings indicate that the robust estimation accuracy of the ensemble learning model is due to its detection of nonlinear correlations and high-order interactions. Furthermore, our findings demonstrate that the synergy of smartphone photograph-based estimations and AOD-based estimations generates significantly greater spatial coverage of PM2.5 distribution than AOD-based estimations alone, especially in densely populated areas where more smartphone photographs are available.
PM2.5 participates in light scattering, leading to degraded outdoor views, which forms the basis for estimating PM2.5 from photographs. This paper devises an algorithm to estimate PM2.5 concentrations by extracting visual cues and atmospheric indices from a single photograph. While air quality measurements in the context of complex urban scenes are particularly challenging, when only a single atmospheric index or cue is given, each one can reinforce others to yield a more robust estimator. Therefore, we selected an appropriate atmospheric index in various outdoor scenes to identify reasonable cue combinations for measuring PM2.5. A PM2.5 dataset (PhotoPM-daytime) was built and used to evaluate performance and validate efficacy of cue combinations. Furthermore, a city-wide experiment was conducted using photographs crawled from the Internet to demonstrate the applicability of the algorithm in large-area PM2.5 monitoring. Results show that smartphones equipped with the developed method could potentially be used as PM2.5 sensors.
3 December 2021 in IEEE Transactions on Geoscience and Remote Sensing
Airborne particulate matter participates in the scattering of light, thus leading to visibility degradation. The widespread use of smartphones provides an opportunity to determine the level of this degradation from a smartphone photograph perspective by extraction of the image features representing aerosol optical extinction properties (AOEPs). This article presents novel algorithms to measure visibility through the extraction of AOEPs (i.e., the local medium transmission rate and the local medium extinction coefficient) from a single photograph. Among them, the transmission rate is derived based on the refined photograph’s dark channel prior and the extinction coefficient from the improved transmission map and depth map. Furthermore, to address the complexity of urban scenes, we draw on visual features manifested on a photograph in various weather conditions, analyze the influence of scene structures on extracted features, and identify the combination of extracted features to improve the estimated AOEPs. We also exploit the suitability of these two AOEPs and develop a method to automatically select an appropriate property from them based on the scene structure of a given photograph. The proposed algorithm is first validated through experiments using public databases. Then, two experiments, one on a city scale and the other on a national scale, are conducted using smartphone photographs crawled from the Internet to evaluate the accuracy of visibility estimation. Experimental results show that the proposed algorithms can estimate the atmospheric visibility in real time; therefore, this study provides an effective method of monitoring environments with crowdsourced data.
COVID-19 resurgences worldwide have posed significant challenges to the formulation of preventive interventions, especially given that the effects of physical distancing and upcoming vaccines on reducing susceptible social contacts and eventually halting transmission are still unclear. Using anonymized mobile geolocation data in China, we devised a mobility-associated social contact index to quantify the impact of both physical distancing and vaccination measures in a unified way such that the gap between intervention measures and disease transmission can be explicitly bridged. This index explained 90% of the variance in the changing reproduction number of infections across the COVID-19 outbreak in Wuhan, and was validated in six other cities of different population densities. Our simulations showed that vaccination combined with physical distancing can contain resurgences without relying on mobility reduction, whereas a gradual vaccination process alone cannot achieve this. Further, for cities with medium-population density, vaccination can shorten the duration of physical distancing by 36%-78%, whereas for cities with high-population density, infection numbers can well be controlled through moderate physical distancing. These findings provide guidance on tailoring and implementing comprehensive interventions for cities with varying population densities.