Excursion incidents pose a significant concern for the aviation sector as they account for most accidents on runways. Accurately predicting tire-pavement friction during landing is crucial to ensure safe operations and prevent runway excursions. However, traditional methods and mathematical models used for measuring tire-pavement friction parameters rely on specialized equipment and expertise, resulting in substantial runway closure time. To address this, the study explores the potential of leveraging pattern recognition and image processing technologies, coupled with the availability of high-resolution and periodic satellite images, to infer tire-pavement friction conditions on runways. This research specifically investigates the determination of the coefficient of friction on runways using aerial images. Two methodologies are compared: one utilizes Convolutional Neural Networks (CNN), while the other utilizes image color composition characteristics (RGB-HSV). The aim is to establish a correlation between these methodologies and coefficient of friction data, enabling the prediction of this physical parameter based on the image data. Data from a Brazilian aerodrome spanning the period of 2015 to 2020, along with aerial satellite images from the same timeframe sourced from the Google Earth Pro platform, are employed. The evaluation reveals promising results, with both techniques achieving a mean squared error of less than 0.01. The absolute errors are in the range of 0.1 for the RGB-HSV technique and 0.08 for the CNN technique. These findings highlight the potential of using aerial images and advanced image analysis techniques to accurately estimate tire-pavement friction conditions on runways, providing valuable insights for enhanced safety measures and decision-making in aviation operations.