The state of the atmospheric environment is related to the values of atmospheric path radiation in remote sensing data.As an important information resource

an atmospheric path radiation image derived from remotely sensed data can be used in many research fields.These include the research in the transmission properties of the atmosphere

the atmospheric modification of remote sensing image

the study of both regional and urban atmospheric environment

and the monitoring of the atmospheric environmental qualities.The remote sensing data collected by earth observation technology includes predominantly ground information and weak atmospheric information.The atmospheric path radiation remote sensing image which includes atmospheric information can be generated by separating the weak atmospheric signals from the strong ground returns in the remote sensing data.The key restriction on atmospheric environmental remote sensing lies both on the above process and extracting atmospheric information from the remote sensing data.Therefore

it is important to derive the data of quantitative atmospheric environmental pollution and precise atmospheric modification of the remotely sensed digital image.After many years of previous research

a scheme of generating the image of atmospheric path radiation by computing the value of atmospheric path radiation of each pixel using land reflectance is demonstrated in this paper.The principle and methodology for generating the heterogeneous atmospheric path radiation image are based on known surface reflectance.The atmospheric path radiation remote sensing image with different waveband from multi-spectral or hyper-spectral remote sensing image can monitor urban atmospheric pollution with different type and degree（size and content of aerosol） without ground interference.In this paper

the research of urban atmospheric pollution monitoring is developed from Shanghai City as a demonstration

with using the atmospheric path radiation remote sensing images from MODIS multi-spectral remote sensing image

based on the atmospheric pollution observational data（PM10 thickness）.Using fuzzy neural network method

the input neuron altogether includes MODIS atmospheric path radiation remote sensing images（4 wave bands）

the traffic density images

and the NDVI grid images.The output is the single neuron of the synthetic evaluation classification images of atmospheric environment quality

and the network membership function is Gaussian function and sigmoid excitation function.According to the synthetic evaluation classification images

Shanghai atmospheric environment quality is basically the first and the second standard classification.In 6 day-long research time intervals

the atmospheric environment quality spatial and temporal pattern has changed obviously.On 3rd

9th and on 10th the atmospheric environment quality are the second level of majorities

and on13th the mass of atmosphere quality achieved the highest level standard

on 21

2

2nd by gradually drops

the pollutant atmospheric environment quality are the second level gradually.During the transmission in the atmosphere

electromagnetic wave is influenced by the atmospheric path radiation（that is

the upwards dispersion along the optical path） which is related to atmospheric environmental quality parameters such as the atmospheric aerosol optical depth

the total absorbing particulates and contents of the gaseous pollutants.At the same time

the atmospheric path radiation is very useful information in the remote sensing retrieval of regional atmospheric environmental qualities.It is also a supplement to the technology development of traditional atmospheric environmental remote sensing.In the future

we would expect this technique to be used in the retrieval of the atmospheric vapor content of the globe from GPS signals.