Research On The Problems Of Edge Detection And Image Denoising Of The Video Image In The Strain Test

With industrial manufacturing technology and processing technology enhancement and improvement, detection methods of detection speed and accuracy has put forward higher requirements, and precision measurement technology is the foundation and prerequisite for industrial development, however, the existing detection methods (such as caliper, microscopes, COMERO) are difficult to balance between speed and accuracy of the contradictions.then image measurement technologies come into being, and be more extensive.By using image technology, combined with video strain measurement system (video-extensometer),deeply we researched on the problem of edge detection and image denoising which affect the accuracy of the video strain measurement system, researched the video precision strain measurement system used to the testing machine. And analyse the existing problems in the current issues, find solutions to the problem.The main contents of this text as follows:1. Based on the ideal optical system imaging principle, established a system imaging model, elaborate on the principle of the video strain measurement system and system calibration method, also with the Modulation transfer function to description of the image quality imaging system. From the atmosphere, cameras and optical lenses to detailed analysis the system,and Identified modulation transfer function of video strain measurement system. Introduced the hardware of measurement system, elaborate on the various configurations, Introduced visualization software functionality framework of the measurement system, And the Function mode of the various Implementation components Module.2. In the processing of digital image capture and transfer, due to the interference of various factors, Image will be subject to all kinds of noise pollution inevitably, sequentially resulting in the decline in image quality, then will affect the edge detection, sub-pixel positioning, and other follow-up treatment. This paper based on the actual situation of the strain measurement system, carried out a detailed study of the various noise which appeard in experiments measuring, Analysed noise sources which existed in the strain measurement system, Established a system noise model, used the appropriate digital image filtering technology to eliminate noise, uesd the airspace filter in the work of eliminating noise. compared of the three filtering method,which is mean filtering, median filtering and a new filtering, By contrast the result of the signal-to-noise of the filtering algorithm and experimental verification,we know that the new filtering method to the strain measurement system is relatively good, Mean filtering and median filtering only effect for a single specific noise. and the new filter can eliminate mixed noise system,be propitious to the follow-up treatment system.3. For the image measurement systems, the most direct way to Improve the measurement accuracy is improve the image sensor resolution,however, this resolution by improving hardware to improve the accuracy of methods is not economical and limited. Using software algorithms to improve the system's precision is an effective method. In this paper, which is not to change the hardware configuration of circumstances, with the edge detection and sub-pixel location technology to achieve the goal of positioning, improved the strain measurement accuracy of the measurement system.In this paper,we deeply research on edge positioning technology of the strain measurement system,deeply explore the the rough and exactness location algorithm and the edge fitting algorithm of specimen edge and the marker's edge. First identified the unreasonable of the original differential positioning rough edge detection algorithm,used the Gauss - Laplace edge detection algorithm to operator edge of the rough location. Then used sub-pixel edge algorithm to be exactness positioning, Comparative analysis of several sub-pixel algorithms, Including Moment Method, First Differential expectations and Interpolation Method, thereinto Bilinear interpolation is the best method.then we make use of the principle of Least Square and Statistics Principle to making the edge fitting. Experimental results show that the approach adopted by the strain measurement system improve the accuracy and stability effectively.4. Analysis of the reasons for displacement of elastic deformation of the strain measurement system. For the single-direction tensile, there will be a partial no rules fluctuations in its flexibility and yield stage in the process of alternating phase, which is not the same degree of volatility when using different approaches. Through experimental comparative analysis we know that Least-squares fitting has less volatile when using frames and frame based on the comparison of the threshold value. When the test conditions shift speed, Change direction or stop it will produce a significant fluctuations.5. Analysis of the unstable factors which Affect the accuracy of the strain measurement system. There are two major factors,one is the environmental factors of the strain measurement system,another is the error existed in the system. Such as temperature changes, random vibration, calibration error, lamp-house of error,operator error and Algorithm for Error, All of these will work a certain impact to the system stability and the improvement the system accuracy. this paper discusses these issues deeply and suggests some modifications comments.This paper just seeks to use some methods to improve the accuracy of the image strain measurement system, combines the video strain measurement system to conducted a study on the edge detection and image denoising, which has a certain relative, hoped to play a little help to the development and application of image technology.