Enhancing textural detail in infrared simulation image: a time-segment temperature field modulation method

To improve textural detail in infrared simulation image, a novel land surface infrared radiance image enhancement method with a time-segment temperature field modulation is proposed. It significantly enriches the detail and realism of the infrared simulation image. Initially, the time dependent land heat transfer theory is employed to accurately calculate the surface temperature with the specific land cover. Subsequently, the impact of surface temperature variations with its shortwave absorption under the given environmental conditions is thoroughly analyzed, which confirms that the land surface shortwave absorptivity is directly proportional to its temperature with the variable rate at the diurnal cycle for both of artificial and natural land cover. The land surface shortwave absorptivity can be derived from its visible image, which implies that it is possible to connect the land surface temperature and its visible remote sensing image. This leads to a semi-empirical formula describing the relation between the slight heterogeneous distribution of surface temperature and the fluctuation of RGB value within the specific land cover. Considering that this heterogeneous distribution also changes with the external environmental variations, the diurnal patterns at different times of the day were quantitatively established to characterize how it evolves over time. Utilizing a remote sensing visible image, the land surface temperature field can be modulated pixel by pixel to enhance the realism and reliability of infrared simulation scenarios. The effectiveness of the proposed method is validated by the experiment. It demonstrates in the simulation scenarios, the vegetation at different heights in the grassland shows notable mottling, and the textures resulting from aging on concrete surfaces are distinctly visual, the variance of their temperatures increases from 0 to 13.857 and 80.378 at noon, and from 0 to 5.05 and 39.322 in the evening. The generated land infrared images are richer in detail, providing the strong application prospects for the virtual testbed evaluation of the infrared system.