Rover Relocalization for Mars Sample Return by Virtual Template Synthesis and Matching

We consider the problem of rover relocalization in context notional Mars Sample Return campaign. In this campaign, a (R1) needs to be capable autonomously navigating and localizing itself within an area approximately 50 ×50 m using reference images collected years earlier by another (R0). propose visual localizer that exhibits robustness relatively barren terrain we expect find relevant areas, large lighting viewpoint differences between R0 R1. The synthesizes partial renderings mesh built from matches those R1 images. evaluate our method on dataset totaling 2160 covering range expected environmental conditions (terrain, lighting, approach angle). Experimental results show effectiveness approach. This work informs campaign choice site where Perseverance (R0) will place set sample tubes for future retrieval (R1).