Spacecraft relative navigation method based on generative reconstructed images

Abstract

The invention discloses a spacecraft relative navigation method based on generative reconstructed images. Because the sunlight in the aerospace environment of a spacecraft is parallel light, the lightradiation area and shade area of a target satellite under the sunlight radiation condition are strongly contrasted, and the target image obtained by a vision camera is not complete. The spacecraft relative navigation method comprises the following steps of using the information of a target local image obtained by the vision camera as input quantity; according to the preset satellite attributes, obtaining feature vectors through learning by a deep neural network; reconstructing the local image obtained by the vision camera into a complete image by a generative adversarial method; performing stereo solving on the generated complete image, so as to obtain relative position postures; converting the relative quantity into a spherical coordinate system as the measuring value of a navigation system; using a relative motion equation under the spherical coordinate system as a navigation state equation, and adopting a special filtration algorithm to complete the solving of the relative navigation. The spacecraft relative navigation method has the advantage that the relative navigation can be completed by passive imaging devices of vision camera and the like.

Description

technical field [0001] The invention belongs to the technical field of spacecraft navigation, in particular to relative measurement and navigation technology, in particular to a relative navigation method for spacecraft based on generative reconstructed images. technical background [0002] Visual navigation is an important way of relative measurement and navigation between spacecraft. Since sunlight is parallel light in the space environment where the spacecraft is located, the contrast between the illuminated area and the shadowed area of ​​the target satellite is strong when it is irradiated by sunlight. It is easy to produce overexposure; on the contrary, the part that is not illuminated is in the dark and cannot be imaged. As a result, satellite imagery of targets often produces a sharp divide between light and dark, which is only partially visible. How to solve this problem has become one of the research hotspots in the current field. [0003] The current existing m...

AI Technical Summary

Problems solved by technology

[0007] Further, in view of the incomplete target image obtained by the visual camera in the space environment, a generative confrontation method is used to restore the complete image from the partial image for relative navigation;

Images