A Spatial Zero-Prior Target Acquisition Method Based on Multi-source Visual Information Fusion

Abstract

The invention provides a space zero-prior target acquisition method based on multi-source visual information fusion. According to the space zero-prior target acquisition method, aimed at a space zero-prior information target, measurement for local forms and postures of the integral target and grabbable parts can be achieved through binocular vision recognition, matching and three-dimensional reconstruction methods; multi-source information fusion treatment for position and posture information of the grabbable parts of the target is achieved through a compound measurement method of fixed position binocular vision and motion position monocular vision, accuracy and reliability of information are improved, and input is provided for autonomous motion planning of a mechanical arm; and real-timedesign, optimization and updating of an acquisition path in a reachable range of the mechanical arm are achieved through a multi-constraint optimization search algorithm and on-line autonomous motionplanning, and input is provided for capture of the mechanical arm.

Description

technical field [0001] The invention relates to a space zero prior target capture method based on multi-source visual information fusion, and belongs to the technical field of space debris removal. Background technique [0002] With the gradual deepening of human development of space resources and the expansion of space applications, the number of space debris continues to increase. Space debris poses a great threat to human space resource development activities, and we must find ways to control and slow down space debris. At present, manipulators capture space debris and other zero prior information is a relatively mainstream method of active space debris removal at home and abroad. [0003] Such as figure 1 As shown, the process of space target acquisition and tracking is generally divided into phase modulation stage, long-distance rendezvous stage, homing stage, approach stage, capture stage and other stages. Spatial null prior targets are often non-cooperative or even...

AI Technical Summary

Problems solved by technology

[0006] (2) The ones proposed in the literature focus on the principle, but do not involve the algorithm of target detection and recognition and the related algorithm of manipulator motion planning;

Images