Study On Complex Straight Line Extraction Technology From Aeronautical Composite Material Images

Carbon Fibre Reinforced Plastic(CFRP)is widely used in the field of aviation equipment construction because of its light weight,low coefficient of thermal expansion and corrosion resistance.Due to the anisotropy of the material,it is necessary to lay it accurately according to the preset ply direction to maximize its performance advantage.However,the background of CFRP surface image is complex,the boundary is fuzzy,and there is no obvious line in the image,so it is difficult to obtain the line information from image to confirm the material direction by using the traditional line extraction method.There are some small scattered areas in CFRP texture images,which are visually represented as broken lines.These discrete areas with linear patterns can be defined as complex lines.Combined with the research project requirements of an aircraft industry group in Chengdu and the characteristics of discrete region with line pattern,this paper proposes a line extraction method based on centerline fitting,which realizes the complex line extraction in aviation composite image,so as to obtain the direction information of CFRP.The experimental results show that the proposed method can effectively identify complex linear features in the images of aviation composite materials,and realize the direction detection of composite materials,and its detection accuracy meets the engineering requirements.Aiming at the above problems,the main work of this paper is as follows:(1)In view of the background complexity of aviation composite image,according to the processing effect of different binarization methods,the adaptive binarization method is used to process aviation composite image.At the same time,the influence of each parameter in the adaptive binarization method on the binarization result is analyzed through experiments.The experimental results show that the method can effectively extract the discrete region with line pattern in complex background,and ensure the subsequent work of complex line feature extraction;(2)Aiming at the problem of regional dispersion in aviation composite images,a morphology-based Steger center line extraction algorithm was proposed: the discrete region images were enlarged and connected by morphological method,and the Steger center line extraction algorithm was proposed to identify the line features of the expansion images.Through the experiment of center line extraction,the effectiveness of the proposed algorithm is verified;(3)Aiming at the problem that traditional method cannot extract the features of complex lines from the images of aviation composite materials,a feature extraction method based on centerline fitting was proposed: A method to judge the intersection points based on the connected domain was designed.Based on this,a method to cut and fit the intersection points was proposed.Based on the center line extraction,the line fitting was carried out to extract the line features.Finally,by controlling the shooting conditions and adding the shielding objects to collect the images of different aviation composite materials,experiments are carried out to verify the effectiveness and anti-interference of the proposed algorithm for complex line extraction.The quantitative analysis shows that the Angle deviation of the final extracted line is less than 0.4°,which can fully meet the requirement that the deviation of aviation composite material manufacturing is less than ±3°.