Radiometric cross Calibration of Gaofen-1 WFV Cameras Using Landsat-8 OLI Images: A Simple Image-Based Method

Juan Li,Lian Feng,Xiaoping Pang,Xi Zhao,Weishu Gong

doi:10.3390/rs8050411

Abstract

WFV (Wide Field of View) cameras on-board Gaofen-1 satellite (gaofen means high resolution) provide unparalleled global observations with both high spatial and high temporal resolutions. However, the accuracy of the radiometric calibration remains unknown. Using an improved cross calibration method, the WFV cameras were re-calibrated with well-calibrated Landsat-8 OLI (Operational Land Imager) data as reference. An objective method was proposed to guarantee the homogeneity and sufficient dynamic coverage for calibration sites and reference sensors. The USGS spectral library was used to match the most appropriate hyperspectral data, based on which the spectral band differences between WFV and OLI were adjusted. The TOA (top-of-atmosphere) reflectance of the cross-calibrated WFV agreed very well with that of OLI, with the mean differences between the two sensors less than 5% for most of the reflectance ranges of the four spectral bands, after accounting for the spectral band difference between the two sensors. Given the calibration error of 3% for Landsat-8 OLI TOA reflectance, the uncertainty of the newly-calibrated WFV should be within 8%. The newly generated calibration coefficients established confidence when using Gaofen-1 WFV observations for their further quantitative applications, and the proposed simple cross calibration method here could be easily extended to other operational or planned satellite missions.

Highlights

To build a near-real time, all-weather and global surveillance network for agricultural planning, disaster relief, environment protection and security, the Chinese government approved the implementation of high-definition earth observation system (HDEOS)
With sufficient calibration sites and close estimates of spectral adjustment factors, a linear regression based on Equation (6) could yield cross
The mean relative difference (MD) and root mean the hypothesis of zero gain and offset has been rejected for all the bands, with square root mean difference (RMSD) between the linear relations estimated radiance using digital number (DN) of a significant level of

Summary

Introduction

To build a near-real time, all-weather and global surveillance network for agricultural planning, disaster relief, environment protection and security, the Chinese government approved the implementation of high-definition earth observation system (HDEOS). With high spatial resolution (16 ˆ 24 m at nadir) and wide coverage (4 ˆ 200 km), the four wide field of view (WFV) cameras on-board Gaofen-1 provide detailed observations for the entire globe within four days. The unparalleled information provided by WFV sensors can be used to understand the biological, chemical, and physical processes at both small and large scales, once the relationships. To obtain valid results from remotely sensed data, accurate radiometric calibration is required. The data provider (China Centre for Resources Satellite Data and Application, CCRSDA) published radiometric calibration coefficients on their official website, where in situ spectral measurements of the Dunhuang

Methods

Results

Discussion

Conclusion