Land surface temperature retrieval method for measured data from unmanned aerial vehicle (UAV) mid-wave thermometry thermal imaging cameras (MWTIC)

UAV thermal infrared remote sensing imagery allows for higher resolution LST (land surface temperature) to be acquired, but temperature drift during thermal camera data acquisition reduces the reliability of the data. Data with temperature drift cannot be accurately removed by the camera’s own automatic calibration or by using a fixed calibration function. In addition, during the acquisition of data by a thermal imaging camera at low altitude, the transmission of thermal radiation is affected by the atmosphere at mid-flight altitude, and the data need to be atmospherically corrected to characterize the actual LST. In this paper, the errors caused by temperature drift are removed by feature matching and linear fitting in the data processing process, so as to obtain more accurate mosaics of brightness temperatures. In the retrieval process, the LST is obtained by synchronizing the atmospheric wet temperature contour lines, based on the principle of thermal radiation transmission, and combining with the specific emissivity of the ground to obtain the high accuracy of the LST. The feasibility of the algorithm is verified using continuous actual measured LST. The results show that based on the synchronized atmospheric temperature and humidity contours, the atmospheric influence can be effectively eliminated, and the LST obtained by the retrieval has a high accuracy. From the experimental results, it can be seen that the method proposed and analyzed in this paper is a feasible method to obtain high-precision LST using thermal infrared remote sensing images from UAVs.