The work described in this paper harnesses an open-source astrodynamics engine and visualization to quantify the performance of on-board optical navigation. It introduces the Hough Circles transform as a candidate method for centroid and apparent diameter extraction. The coupled nature of the simulation enables simultaneous pointing and orbit determination with dynamic image generation, all in a realistic flight-software environment. Navigation is done about Mars solely using optical images, and by means of limb or centroid/diameter extraction. Through the implementation of pre-existing algorithms for a baseline comparison and the development of Hough Circles, an end-to-end autonomous flight-software stack is developed and tested. This research provides insight into achievable navigation accuracy and image processing methods, as well as outlier mitigation for mission readiness.

本文描述的工作利用开源的天体动力学引擎和可视化技术来量化机载光学导航的性能。它引入了Hough Circles变换作为质心和视在直径提取的候选方法。仿真的耦合特性使得可以在真实的飞行软件环境中同时进行指向和轨道确定以及动态图像生成。仅使用光学图像并通过肢体或质心/直径提取来完成有关火星的导航。通过实施用于基线比较的预先存在的算法以及开发Hough Circles，开发并测试了端到端的自主飞行软件堆栈。这项研究可洞悉可实现的导航精度和图像处理方法，