Abstract: According to the precise landing of the detector on planet, an algorithm for autonomous extraction and matching of referenced area based on the geometrical characteristics of image was proposed. Firstly, the sample image was preprocessed, including threshold segmentation and connected area extraction. Then, the outline of referenced area's shadow and its linear and planar geometrical characteristics were extracted. Finally, template matching was tested based on Euclid distance with weights. The proposed algorithm could overcome the defect that the selected referenced area was simple.Experiments demonstrate that the accurate rate of this algorithm is higher than 93%, which proves that the proposed method is feasible and robust.