Objective   Thermal infrared remote sensing has the ability of day and night detection and good environmental adaptability, which makes it have important applications in natural ecological environment monitoring, urban heat island effect monitoring, lake and reservoir water quality monitoring, etc. The application of thermal infrared remote sensing has gradually changed from qualitative to quantitative, and absolute radiometric calibration is the prerequisite for the quantification of remote sensing information. Among them, on-board blackbody calibration uses on-board blackbody as the calibration source, which is not limited by time, environment and other factors. It can produce corresponding calibration coefficients for each orbit data, improve the frequency of on-orbit absolute radiometric calibration. Based on the on-board 0-level blackbody calibration data of GF5B VIMI (Hyperspectral observation satellite, visible and infrared multispectral image), the absolute radiometric calibration research of satellite thermal infrared channel is carried out. In this way, reliable calibration results can be obtained to provide a method basis for the subsequent blackbody calibration of satellite thermal infrared remote sensing.

  Methods   Based on the laboratory calibration before the launch of GF5B satellite, the on-board blackbody calibration data is used to establish the on-board absolute radiometric calibration model applicable to the GF5B thermal infrared channel. Firstly, relative radiometric correction is carried out for the high and low temperature blackbody image data transmitted from satellite; based on the blackbody image data after relative radiation correction, the average DN of each channel corresponding to the high and low temperature blackbody is obtained. At the same time, the high and low blackbody temperature is calculated based on the high and low temperature blackbody auxiliary data, and then the radiance value of the corresponding channel of the high and low temperature blackbody is calculated using the Planck function. Then, according to the actual response average DN of the high and low temperature blackbody image and the corresponding channel radiance, the inner blackbody absolute radiometric calibration coefficient is calculated. Finally, the internal and external blackbody calibration conversion coefficients are used to convert the internal calibration coefficients into the absolute radiometric calibration coefficients of the on-board blackbody (Fig.1). In addition, according to the error sources of the on-board calibration system, various indicators affecting the accuracy of the on-board calibration system are analyzed. The accuracy of on-board blackbody calibration is evaluated and verified by using the ground synchronous buoy measurement data.

  Results and Discussions  The on-board blackbody calibration data of the 1 850th orbit on January 12, 2022 are selected to conduct the on-board blackbody absolute radiometric calibration, and its on-board absolute radiometric calibration coefficient (Tab.3) is obtained. Through the analysis of various indicators affecting the accuracy of the on-board radiometric calibration system, the results show that the total error of the on-board radiometric calibration of the camera is 1.268% (Tab.4), and the equivalent temperature is 299.1 K@300 K. Therefore, the absolute calibration accuracy of the on-board calibration system is 0.9 K. The verification results of satellite-ground synchronization show that the relative differences of radiance of B11 and B12 channels are 0.64% and 1.35% respectively. The brightness temperatures of B11 channel monitored by satellite and ground measurements are 273.78 K and 273.45 K respectively, with a difference of 0.33 K; the brightness temperatures of B12 channel monitored by satellite and ground measurements are 272.58 K and 273.35 K respectively (Tab.5), with a difference of 0.77 K, which shows that the brightness temperature difference is within 0.8 K. The satellite-ground data have a good consistency, which indicates that the thermal infrared channel of GF5B satellite has a high calibration accuracy on orbit, and the results are true and reliable.

  Conclusions   The on-board blackbody calibration method based on GF5B thermal infrared channel has good accuracy and reliable calibration results, which can meet the needs of remote sensing data quantification application. It provides a method reference for real-time and accurate acquisition of thermal infrared channel calibration coefficient. The construction of the research method is based on GF5B on-board calibration blackbody, which has important reference value for the on-board blackbody calibration of other satellites.