Data acquisition method based on deep learning and multi-eye vision in digital twin environment

Abstract

The invention discloses a data acquisition method based on deep learning and multi-eye vision in a digital twin environment. The steps are as follows: S1, setting a spherical marker point that is distinguishable from the environmental background; S2, obtaining the spherical shape of the marker point in the video image Coordinates and radius of the center position; S3, build a deep learning model and perform training; S4, attach the marker points to the target object with positioning, use step S3 to locate the marker points in space, and realize the positioning of the target object. The invention can be used to acquire the position and attitude data of various target objects in the digital twin environment, and has strong universal applicability. With the aid of landmarks, the complexity of visual image analysis and processing is reduced, making the identification and positioning process simpler, more efficient and more reliable. Use deep learning to locate the marker points, minimize the positioning error of the camera due to image distortion, and can adapt to various camera numbers and layouts.

Description

technical field [0001] The invention belongs to the technical field of digital acquisition, and in particular relates to a data acquisition method in a digital twin environment, especially a data acquisition method based on deep learning and multi-eye vision. Background technique [0002] Digital twin technology requires a high degree of simulation of physical equipment, and at the same time needs to grasp various state data of physical equipment in real time, so that the simulation model is consistent with the real-time state of physical equipment. Digital twin technology relies on perception and control technology and its comprehensive technology integration. The mechanical state information, current state information, thermodynamic state information and action state information of physical equipment need to be obtained by means of sensing technology. [0003] In the construction of the digital twin system, it is first necessary to construct the 3D model of the physical e...

AI Technical Summary

Problems solved by technology

[0010] According to the above-mentioned prior art, the technical problem to be solved by the present invention is that the installation of sensor data collection equipment is complicated and not universal, the algorithm of directly identifying and locating target objects using machine vision methods is complex, and the algorithm of indirect identifying and locating target objects is indirect by means of marker points When using the deep learning algorithm to recognize the unstable problem when the background changes, and the positioning error problem caused by the calibration complexity and image distortion when positioning the target through binocular or multi-eye vision, thus providing a digital twin environment based on deep learning and multi-eye Visual Data Acquisition Methods

Images