Autofocus

RANSAC

Physics

Trajectory

Mathematics

Outlier

Algorithm

Image (mathematics)

Optics

Astronomy

Inertial navigation system

Orientation (vector space)

Artificial intelligence

Azimuth

Focus (optics)

Geometry

Computer science

Synthetic aperture radar

Computer vision

IEEE Geoscience and Remote Sensing Letters

Unmanned aerial vehicle (UAV) synthetic aperture radar (SAR) is usually sensitive to trajectory deviations that cause severe motion error in the recorded data. Because of the small size of the UAV, it is difficult to carry a high-accuracy inertial navigation system. Therefore, in order to obtain a precise SAR imagery, autofocus algorithms, such as phase gradient autofocus (PGA) method and map-drift (MD) algorithm, were proposed to compensate the motion error based on the received signal, but most of them worked on range-invariant motion error and abundant prominent scatterers. In this letter, an improved MD algorithm is proposed to compensate the range-variant motion error compared to the existed MD algorithm. In this context, in order to solve the outliers caused by homogeneous scenes or absent prominent scatterers, a random sample consensus (RANSAC) algorithm is employed to mitigate the influence resulting from the outliers, realizing robust performance for different cases. Finally, real SAR data are applied to demonstrate the effectiveness of the proposed method.

An Improved Map-Drift Algorithm for Unmanned Aerial Vehicle SAR Imaging