Feature extraction and image registration in the final approaching phase of rendezvous and docking

Motion information of the target spacecraft can be calculated according coordinate change of feature points based on the motion control model. In order to extract motion information including velocity, location, attitude of the target spacecraft in the final approaching phase of rendezvous and docking, feature extraction and image registration for the target spacecraft is an essential step. In this paper, the original images are obtained by CCD. Circle Hough Transform based on wavelet adaptive thresholding denoising method is utilized to extract the edges and the characteristic circles of the target spacecraft. Then we designed three feature points on the fore-end cabin. Finally, images registration based on control points is applied to the reference and to be register images. The results of images registration demonstrate that three feature points on the fore-end cabin of the target spacecraft model is especially effective. Additionally, experimental results shows that running speed is improved greatly and high accuracy is provided for circles detection. It will directly affect further extraction of motion information, so there is no doubt that is a vital step.