This study investigates the application of Earth-based crater remote sensing recognition methods to Martian surface imagery. To address the transferability limitations of traditional image processing techniques in planetary crater detection

we designed a workflow that integrates multi-scale Hough transform with physical indicators such as ring completeness and radial gradient. We verify the effectiveness of the approach through experiments on Earth remote sensing imagery: the method reliably detects primary craters with small radius estimation errors

while candidate generation is identified as the major bottleneck due to its high computational cost. The analysis further reveals that detection performance is strongly influenced by illumination conditions and edge sharpness

with limited sensitivity to low-contrast crater rims. Finally

we discuss the potential value of incorporating geophysical indicators and multi-source data in future Martian crater recognition. Overall

this lightweight

training-free method proves feasible in resource-constrained scenarios and offers a valuable reference for visual navigation in planetary exploration.