A method and application of precise haze traceability based on image recognition

Abstract

Haze affects the urban landscape, increases the cost of cleaning the city, and directly or indirectly causes respiratory diseases and affects the physical and mental health of citizens. In addition, the complex causes of haze, the difficulty of tracing the source leads to the current "wide net" type of treatment measures have limited effect, haze tracing work there are technical bottlenecks, poorly targeted treatment programs. Some economic and technological backwardness but also more serious pollution of the township, and even the use of "shock therapy" to manage the haze, although effective, but often cause economic decline. In this regard, the article proposes an accurate traceability method for haze, which achieves accurate traceability of haze through highly reductive haze collection, multi-angle morphological analysis and haze source database, and in addition applies remote sensing technology to obtain the real particle size parameters and temporal-spatial distribution of haze by inversion of AOD-PM2.5 mathematical model to check the obtained results from a macroscopic perspective, which ensures more The accuracy of the data is ensured. The results show that the pollution sources of typical haze in Xi'an are, in order of contribution, industrial coal combustion (35.1%), automobile exhaust (26.0%), industrial smelting (13.7%), soil sand (11.5%), and mineral extraction (1.9%). Among them, industrial coal combustion, automobile exhaust, and industrial smelting are the main sources of typical haze, with a combined contribution of more than 60%. To summarize the experience gained during the study, pollution prevention and control recommendations for industrial coal combustion, automobile exhaust and industrial smelting are proposed.