Optical remote sensing has limitations in obtaining images due to weather and environmental effects, so these limitations must be overcome to produce time-series image data. As an alternative to this, research are being conducted to simulate images at a specific time for which a specific image is needed. The purpose of this study is to improve the results of this process by preprocessing the input images of a multiple linear regression model alongside other remote sensing image simulation methods. Specifically, the input images, which are applied to a multi-linear regression equation, are preprocessed for phenological and radiometric normalization by a random forest regression model. The experimental results show that the proposed method is superior to the conventional methods both visually and quantitatively.

由于天气和环境的影响，光学遥感在获取图像方面有局限性，因此必须克服这些局限性才能生成时序图像数据。作为对此的替代，正在进行研究以在需要特定图像的特定时间模拟图像。这项研究的目的是通过预处理多元线性回归模型的输入图像以及其他遥感图像模拟方法来改善此过程的结果。具体而言，对输入到多线性回归方程的图像进行预处理，以通过随机森林回归模型进行物候和辐射归一化。实验结果表明，该方法在视觉和定量上均优于常规方法。