Mid-Infrared Global Surface Radiation Information Generation Method Research Based On MODIS Data

With the rapid development of science and technology,more and more scientific researches have been carried out for change detection and characteristics analysis of land surface on global scale.Therefore,the demand for radiation data on global scale has gradually increased.MODIS products have become the main data source for global scale research with their short revisiting period,wide swath,and full spectral imaging capabilities.Among them,the mid-infrared(MIR)bands are valuable in applications such fire disaster monitoring because of their high recognition of high-temperature targets.In addition,during the operation of aircrafts such as planes and missiles,their tail flame of combustion material has strong absorptivity in MIR bands.Therefore,the MIR bands can be used for target detection and tracking.However,due to the strong absorption of atmospheric molecules in the MIR bands,it is difficult for the background radiation signal to pass through the atmosphere and reach to the sensor.Therefore,it is difficult to directly obtain the background radiation characteristics of MIR bands,which could be synthesized by the image simulation technology..Aim to the difficulties in obtaining full-channel continuous background radiation data in MIR bands,this study conducted a research on the generation method of global background radiation data in MIR bands based on the MODIS data product.The full-channel continuous surface emissivity data and surface temperature data at global scale are generated with follows:First,for the mid-infrared wave band emissivity data generation,the global land surface reflectivity(LSR)data in the visible near-infrared region was synthesized.Due to factors such as clouds and deep sea in MODIS data acquisition,missing and invalid pixels may exist on the image and the problem was addressed.Second,in order to generate full-channel continuous surface emissivity data in MIR bands,the method based on spectral correlation of adjacent bands was used.The source bands were selected from MODIS short-wave infrared reflectivity data,and the regression models between source bands target emissivity bands were constructed.Full-channel continuous background radiation image in MIR bands were simulated with the models.The error of models was less than 5%,except that the relative average error at 2.7?m was9.6%.Third,after the simulation of emissivity image,the global land surface temperature(LST)image was generated based on MODIS LST and Sea Surface Temperature(SST)data products.Data product was first repaired,then,the data attributes of land surface temperature data and sea surface temperature were unified.Global scale data synthetic processes and methods were proposed.Conclusions were made based on the analysis of the synthesized global background radiation data.On the one hand,the repairing method addressed the problem of invalid and missing data efficiently.And the realization of automatic procession flow for global scale data generation could build a good data foundation for related research at the global scale;On the other hand,the emissivity conversion models based on spectral correlation has good simulation accuracy,and the established models effectively solved the difficulties in acquisition of full-channel emissivity data in MIR bands.It can provide theoretical support for applications such as image simulation and sensor development in MIR bands.