Research On Thread Vision Measurement Technology Based On Image Fusion

Modern industry not only requires high precision requirements for yarn detection,but also needs to achieve high efficiency in detection efficiency to meet modern highly efficient industrial requirements.Thread contact measurement is often inaccurate and thread quality cannot be monitored in real time.When a thread is detected by image processing,it is impossible to obtain a clear thread edge due to the spiral angle.Enter multi-focus fusing of floss images and perform fusion processing for multiple floss images with distortion of the tooth profile to obtain a floss image with sufficiently clear edges.When examining the evaluation of the yarn image sharpness,the yarn edge image is innovatively used to calculate the evaluation value,and the obtained evaluation value can effectively reflect the sharpness of the yarn fusing image.Finally,the buckling angle of the fused image and the unfused image is experimentally calculated for comparison and analysis.The main research content of this thesis is divided into the following four parts:(1)According to the property of the thread profile,the image of the points on the edge of the thread profile are analyzed and the model of Edge imaging is established to analyze causes the distortion of the thread profile.in the meantime,to build up the quality of the thread profile,select an available image pretreatment method.(2)The automatic registration and image fusion methods of thread images are studied.After multiscale thread images with Blur edges are processed by the thread image fusion algorithm,a thread profile image with clear edges is obtained.(3)Investigate the image sharpness evaluation method and find the thread image quality evaluation method.It mainly evaluates the image quality of the yarn after the fusion and calculates the evaluation value of the yarn image before and after the fusion through experiments.(4)The thesis studies the method of Dental Angle parameters of the thread image profile after fusion.The sub-pixel positioning method is first used toextract the edge of the thread,and the flank angle of the thread iscalculated before and after fusion.through experiments,and at the same time comparative experiments are carried out.It is drived that the error of the profile angle of the image of the thread after the fusion is smaller.