Vibration And Displacement Field Measurement Via High-Speed Imaging System

Measurement technology is one of the most basic technologies in modern industrial technology.The accuracy and efficiency of this technology often determine the prospects and the upper limit of the development of the manufacturing industry.With the progress of manufacturing techniques and processing methods,the higher accuracy and faster speed of the measurement techniques are desperately in need,and traditional contact measurement methods have been unable to meet the measurement needs in many cases.Since the fast development of image acquisition systems and computer image processing technologies,vision-based image measurement methods have become a new high-performance measurement technology.More and more attention has been paid on this newly developed measurement approach due to its advantages of non-contact,high-precision,full-field measurement of image,etc.The vision-based measurement technology is based on optical imaging systems and incorporates modern techniques such as computer graphics,image recognition and tracking,and optoelectronics to obtain the measurement information through the processing and analysis of two-dimensional images.Nowadays image measurement technology is widely used in various fields such as intelligent transportation,unmanned factories,and product quality inspection,etc.In this paper,a high-speed camera system is used as a tool to deeply study the measurement system and the measurement algorithms,and a series of fast,reliable single-point displacement extraction algorithms and full-field displacement measurement methods are proposed in different conditions.This paper first introduces the background and significance of image measurement techniques.And then we introduce the high-speed camera system used in this paper in details,and analyzes the applicable scenarios under different hardware.Also we derive the principles of optical imaging and the correction of optical distortion.The commonly used single-target displacement extraction algorithms based on template matching strategy are briefly explained.The algorithms are improved to satisfy the engineering applications to achieve the measurement accuracy and real-time performance.We have developed a software using these algorithms for the extraction of vibrations from the videos captured by the high-speed image system.It has achieved high accuracy and real-time performance in experimental applications.This paper uses the vibration measurement of the column impact test and the forklift steering wheel system as examples to verify the effectiveness of the measurement system in practical applications.Then a full-field displacement measurement method based on digital image correlation(DIC)method has been introduced.The traditional digital image correlation algorithm based on Newton-Raphson method has the disadvantages of high resource occupancy and low computational efficiency.In this cases,this paper has proposed an algorithm based on particle swarm optimization to obtain the optimal solution for the correlation function and improve the performance of traditional DIC method.Intensity-gradient-based methods have great advantages in the calculation speed for the full-field displacement calculation,but it is sensitive to the illumination change and has a poor calculation accuracy when there is a large displacement.In this paper,full-field displacement measurement methods based on optical flow method have been proposed to enhance the performance of the intensity-gradient-based methods.These improvements include using the most reliable search strategy to acquire the initial matching point at sub-pixel accuracy quickly and accurately,emphasizing the weights of the desired points and using weighted-least-square method instead of the traditional least squares method for solving linear equations,using iterative optical flow method,etc.,which greatly improves the calculation accuracy and maintains the advantages of the original calculation speed.This paper also uses a multithread calculation based on reliability search to solve the applications with abrupt changes in the displacement field(such as fractures).The proposed algorithms are all verified using the simulated speckle patterns and the DIC Challenge speckle image set.The test results show that the proposed methods have high calculation accuracy and computational efficiency.All results in this research demonstrate that it has great practical value.