Space Object Recognition With Stacking of CoAtNets Using Fusion of RGB and Depth Images

Space situational awareness (SSA) system requires recognition of space objects that are varied in sizes, shapes, and types. The images challenging because several factors such as illumination noise thus make the task complex. Image fusion is an important area image processing for various applications including RGB-D sensor fusion, remote sensing, medical diagnostics, infrared visible fusion. Recently, algorithms have been developed they showed a superior performance to explore more information not available single images. In this paper, we compared methods RGB Depth object classification task. experiments were carried out, was evaluated using 13 metrics. It found guided filter context enhancement (GFCE) outperformed other terms average gradient (8.2593), spatial frequency (28.4114), entropy (6.9486). additionally, due its ability balance between good inference speed (11.41 second), GFCE selected stage before feature extraction stage. outcome method fused used train deep ensemble CoAtNets classify into ten categories. learning bagging, boosting, stacking trained purposes. combination able improve accuracy largely baseline by producing 89 % F1 score %.