Three-dimensional object European space reconstruction measurement system based on vision and active optics fusion

Abstract

The invention discloses a three-dimensional object European space reconstruction measurement system based on vision and active optics fusion. The system comprises a monocular vision module, a photometry range finding module, a preprocessing module, an initial three-dimensional reconstruction module, a SLAM posture calculating module, an accurate 3D reconstruction module and a correlation auxiliary mechanism. Through a new method, a mutual relation of an image sequence acquired from the monocular vision module and single point distance information acquired from a laser range finding module is established, space three-dimensional information of each characteristic point of an observed object is calculated, and through iteration optimization, accurate three-dimensional measurement information of the object and position attitude information of the object are finally acquired. Through adopting combination of a high speed Monocular vision system and a laser range finding system, based on high frame frequency detection, data processing pressure is greatly alleviated and the system is especially suitable for space or underwater motion object observation. The system possesses advantages that an imaging speed is rapid; reconstruction precision is high; and the system is simple, the size, weight and power consumption are small and so on. The system is suitable for applications of space and underwater operation robots.

Description

technical field [0001] The invention belongs to the field of three-dimensional reconstruction of non-cooperative targets, in particular to a measurement system for European space reconstruction of three-dimensional targets based on fusion of vision and active ranging. Background technique [0002] In the measurement tasks of space and underwater non-cooperative targets, the short-distance reconstruction and measurement of targets is the key technology. Since non-cooperatives cannot provide effective cooperative information, it does not have space targets equipped with communication response mechanisms or other active sensors, and other robots cannot identify or locate such targets through electronic interrogation or transmitting signals. When the target's motion state and spatial structure are unknown, the target can be detected and reconstructed in 3D scale through vision and other means, and its pose can be accurately measured to provide conditions for further operations. ...

AI Technical Summary

Problems solved by technology

However, due to the immature technology at present, it is impossible to achieve high-pixel, high-precision three-dimensional target measurement, and its lateral resolution is not ideal, and it is difficult to popularize due to the high cost of equipment and high hardware requirements.

[0007] Some combine monocular vision with lidar, such as Jose Padial of the Department of Astronautics of Stanford University, to improve measurement accuracy. However, due to the existence of cameras and rotating mirror systems, the entire system has a large volume, power consumption, and weight, making it difficult to adapt complex environment application

Images