A Method of Homologue Points Extraction for InSAR Mapping with Sparse Control Points

Automatically extracting the Homologue Points (HPs) is a key step to achieve the automatic mapping of InSAR. Subjecting to constraints of the interferometric phase, a quality-map-oriented approach is proposed to extract HPs from the view of detecting and matching features between adjacent SAR images. This method adopts Scale Invariant Feature Transfor (SIFT) and Speed-Up Robust Feature (SURF) to detect interest points, and then the combination of three steps matching technique with criteria of interferometric optimization attributes to automatic HPs detection. Experiments on the same strip and adjacent strips demonstrate the efficiency and feasibility of the technique in the presence of magnitude speckle including significant difference in the imaging geometry.