Markers Identification for Relative Pose Estimation of an Uncooperative Target

In the past ten years, deep learning (DL) has profoundly influenced the development of computer vision algorithms, enhancing their performance and robustness in various applications like image classification, segmentation, and object tracking. This momentum has carried into spacecraft pose estimation, where DL-based methods have begun to surpass traditional feature-engineering techniques as reported in the literature [1-3], corner and marker detection algorithms such as Shi-Tomasi, Hough Transform methods [4, 5]. CNNs have the edge over feature-based methods primarily due to their enhanced robustness against poor lighting conditions and their streamlined computational demands. However, when it comes to space imagery, the scenario changes due to the distinct challenges such as high contrast, low signal-to-noise ratio, and inferior sensor resolution, which can diminish accuracy.