Promising Design Of Satellite Detecting and Tracking Optical system Based On Deep Learning

Abstract

In the past few decades, there has been an urgent need for satellite detection and tracking systems due to the presence of dense space debris around the Earth and to avoid collision of space debris with satellites. Which is mainly applied in space awareness systems, Therefore, this research presents an optical system based on the Mach-Zehnder interferometer for its ability to separate wave spectra with high accuracy and to help outer space image analysis programs such as the TYCHO program to detect and analyze the captured Astronomy image in real time. The proposed optical system will be designed, implemented, installed with a telescope and tested in one of the observatories in (NRIAG), Cairo, Egypt. The results showed the average classification accuracy has been established at 98.9%, compared with Classical machine learning approaches demonstrate accuracy even with a 10%/90% training data split. The TYCHO program Notwithstanding their effectiveness, support vector machine SVM and modified simple ratio MSR exhibit sensitivity to the spatial arrangement of pixel data. Altering a pixel's image position will significantly impact performance. Principal component analysis integrates visual data with geometry information. Owing to their superior visual processing capabilities, they can discern the visual attributes of celestial objects. Convolutional Neural Networks can accommodate variations in illumination and object distance during training. Every approach shows strong performance on the testing dataset. This dataset depicts conditions of space lighting, with all celestial objects documented against a consistent black background, resulting in some variability in magnitude (range) and orientation data.