Application Of Power Equipment Maintenance Based On Markerless Augmented Reality Technology

Augmented reality technology is a popular technology in recent years.Compared with virtual reality technology,augmented reality technology better integrates the virtual environment and the real environment,improving the immersion of the user,and also highlighting the man-machine.Interactive friendliness.In recent years,with the advancement of science and technology,the application scenarios of augmented reality technology have become more extensive,and their accuracy,real-time and robustness requirements have become higher and higher.As the core technology of augmented reality technology,3D registration technology takes a large proportion of time and calculation in the augmented reality system,and there is a cumulative error in tracking registration.In order to avoid the large amount of calculation and poor real-time performance in 3D registration And the problem of error accumulation,this paper proposes a hybrid algorithm-based non-identified tracking registration method and applies it to power equipment maintenance to verify the effectiveness of the method.The following aspects are mainly studied:(1)The basic theory of three-dimensional registration is studied.The two stages of feature point extraction,matching and tracking are analyzed.The advantages and disadvantages of several different feature point extraction and tracking algorithms are compared.At the same time,the needs for equipment maintenance are studied.Firstly,this paper studies the basic theory of 3D registration,analyzes the two stages of feature point extraction and matching,compares the advantages and disadvantages of several different feature point extraction and tracking algorithms.At the same time,the requirement in equipment maintenance is studied.(2)After comprehensive analysis,a hybrid feature matching algorithm was proposed for the extraction and matching of feature points.The shortcomings of single feature matching algorithm,such as low accuracy and time consuming,were solved,and the algorithm was optimized and improved.This algorithm uses FAST algorithm to quickly extract a large number of feature points,and then uses SURF algorithm to describe the extracted feature points,so that it has rotation invariance and scale invariance,so as to realize the comparison and recognition of real images and template images.(3)Aiming at the tracking problem of the target object,a tracking algorithm is designed to reduce errors and eliminate errors.This algorithm improves the traditional LK optical flow pyramid tracking algorithm,adds a diamond search template,improves the search efficiency while reducing the iteration error of the pyramid,and removes the wrong points by the median flow algorithm,thereby achieving the purpose of tracking moving objects,and The fault information display in the maintenance of power equipment is realized.Taking transformer as an application example,simulation experiments are performed to verify the effectiveness of the hybrid algorithm for power equipment maintenance based on unidentified augmented reality technology.Simulation results show that the proposed hybrid matching algorithm can improve the stability and efficiency of the recognition process when the size and angle of the model to be matched change.At the same time,the proposed tracking algorithm to reduce errors and eliminate errors can improve the accuracy of virtual-real fusion.The hybrid algorithm can improve the tracking registration rate and robustness under large displacement conditions.