An Efficient Geometric Rectification Method for Planetary Linear Pushbroom Images Based on Fast Back Projection Algorithm

Planetary mapping products are essential to landing sites selection and scientific research such as geology and geomorphology. Planetary remote sensing images are mainly acquired by linear cameras. Existing geometric rectification software tools in planetary mapping community show unsatisfactory computational performance. In this paper, an efficient geometric rectification method for planetary linear pushbroom images is developed based on a fast back projection (or ground-to-image) algorithm. Additionally, a generic rigorous sensor model is introduced to support different planetary linear pushbroom images. The proposed method was tested using images acquired by Lunar Reconnaissance Orbiter Narrow Angle Cameras and Mars Express High Resolution Stereo Camera. The geometric accuracy was verified using the orthophotos generated by Integrated Software for Imagers and Spectrometers (ISIS). The root mean square errors of conjugate points on the generated orthophotos were less than 0.3 pixels. In case of single-threaded programming, the proposed geometric rectification method was about 4 times faster than ISIS. Whereas multi-threaded programming with four worker threads was more than 3 times faster than single-threaded programming. The experimental results demonstrated that the proposed method significantly improved the computational performance of geometric rectification for planetary images.