A robust photo‐based PM2.5$_{2.5}$ monitoring method by combining linear and non‐linear learning.

Good health is pursued by people all over the world. However, the continual industrialisation has led to more and more atmospheric contamination, and PM2.5$_{2.5}$ has caused serious harm to our life safety and living environment. Without increasing the cost of sustainable industrial production, more and more attention has been paid to the related researches on improving PM2.5$_{2.5}$ monitoring, prevention and control level. Therefore, it is extremely urgent to establish a robust PM2.5$_{2.5}$ monitoring model that can adapt to a variety of scenarios, not only in local places like campuses but also in wide area like city. Existing work has proven that PM2.5$_{2.5}$ monitoring can be achieved by means of photos. But experiments show that the stated‐of‐the‐art methods are far from ideal for PM2.5$_{2.5}$ monitoring when the author tested the performance in two public datasets. To solve the aforesaid issue, this paper ulteriorly proposes a novel photo‐based PM2.5$_{2.5}$ monitoring model, which fuses the results of existing methods by firstly using the weighted average based on the least absolute shrinkage and selection operator regression for learning the basic linear component, secondly using the support vector regression for learning the non‐linear residual component, and finally incorporating the above two outputs to infer the final PM2.5$_{2.5}$ concentration. The main contributions and innovations of this paper are embodied in: (1) the innovative use of image quality assessment model to extract 9 features for PM2.5$_{2.5}$ monitoring, (2) separately extract macro information and micro information from PM2.5 pictures, (3) two newly‐established large‐scaled datasets are applied to verify the effectiveness and robustness of the proposed PM2.5$_{2.5}$ monitoring model. Experiments show that on the latest PM2.5$_{2.5}$ datasets (local and wide), the proposed model has achieved high performance and demonstrated strong robustness. [ABSTRACT FROM AUTHOR]