The Research Of Fibre Orientation Measurement For Carbon Fibre Weave Based On Machine Vision

In the past few decades,advanced carbon fiber reinforced composite materials(Carbon Fiber Reinforced Polymer,CFRP)have triggered a revolutionary transformation in material science.Among them,the most widely used ones are undoubtedly high specific strength,high specific modulus,and fatigue.Carbon fiber composite materials with excellent characteristics,heat resistance and corrosion resistance,etc.,and the tow angle in the carbon fiber composite material has a great influence on its mechanical properties.With the gradual increase in the demand for composite material production in the industry,it is increasingly necessary to rely on machine vision to obtain the angle of carbon fiber tow.Aiming at the difficult problem of tow angle detection caused by the complex optical properties of carbon fiber materials,which are prone to specular reflection and high light absorption,this paper establishes three machine vision-based carbon fiber tow angle detection methods,namely:(1)According to the light source reflection principle of photometric stereo vision,an angle detection method based on photometric stereo vision is established.The traditional photometric stereo vision method calculates the directional gradient of the object surface according to multiple images taken under different light source directions.This article applies it to the calculation of the tow direction of carbon fiber fabric: first obtain the carbon fiber fabric under different illumination angles.Reflect the image,pre-process the image and divide it in equal proportions,calculate the angle of the carbon fiber tow per unit area,and then integrate and analyze all the calculation results.The clustering algorithm is used to classify the detection results,and a pixel map is established to describe the tow angle of the carbon fiber fabric.(2)According to the specular reflection phenomenon of carbon fiber fabric under certain illumination angle conditions,the gray value difference between warp and weft yarns is obtained,and an angle detection method based on edge detection and image fusion is established.Looking for images of carbon fiber fabrics in which only one type of mirror reflection occurs between the warp and weft yarns under a specific light source angle,the target image information is integrated through the image fusion algorithm,so that the warp and weft yarns have a larger gray value difference.The Boolean threshold algorithm is used to highlight the edge information of the target object,the edge detection algorithm extracts the characteristics of the target object,the connected domain algorithm selects qualified target objects,and the minimum area circumscribed parallelogram method assists in the calculation of the tow angle.In view of the large noise density in the original image,a mean filter to suppress noise is added,and steps such as gray value adjustment are added to the effect of diffuse reflection that affects the fusion image,which improves the accuracy of angle detection.(3)According to the principle that the spectrum image of the Fourier transform image will rotate the corresponding angle in the same way with the rotation of the original image,an angle detection method based on the Fourier transform is established.The frequency component intensity of each azimuth angle is obtained from the Fourier transform spectrogram of the carbon fiber fabric image,and the carbon fiber tow angle is calculated according to the frequency component intensity map,and the anti-noise ability and robustness of the method are verified.After completing the research and programming of the three detection methods,the accuracy experiments of the three detection methods have been carried out.The experimental results show that they are suitable for different weaving methods(unidirectional fabric,plain weave fabric,two-dimensional angled fabric and orthogonal fabric)The carbon fiber tow angle detection can ensure stable accuracy,and the detection angle error is within 7°.The efficiency of the three detection methods was evaluated by comparing the detection time,and their different requirements for the working environment were discussed.The results show that the detection method proposed in this paper for the tow angle in carbon fiber fabrics can be effectively applied to other composite fabrics or woven fabrics after adjusting certain parameters.Therefore,the applicable scope of the three detection methods is also analyzed.Finally,the three detection methods are horizontally compared,and the overall analysis and evaluation of the three detection methods are made.