Research On Condom Surface Defect Defection Based On Machine Vision

With the continuous improvement of people's health awareness,the demand for condom is constantly increasing.At present,in the condom production process condom surface defects can only be detected manually,and there are human factors in the detection process,which is time-consuming and laborious.Based on machine vision technology,defect detection is widely used in automation production in all walks of life because of its characteristics of zero contact,fast detection speed and high detection accuracy.In order to realize the automation of condom surface defect detection,this thesis proposes to applying machine vision detection technology to detect of condom surface defects.The main research contents of this thesis are as follows:Firstly,the principle of machine vision detection and the research status at home and abroad were fully understood,and the common defects in condom production were identified.According to the detection requirements of machine vision,the hardware of camera,lens and light source needed by the system was reasonably selected,and on the basis of understanding the common machine vision software,Open CV was selected as the software development platform of the system.Secondly,due to the shape of the die rod itself,there is information compression on both sides of the edge of the condom when the image is acquired.In order to solve this problem,a mathematical modeling study was carried out with a die rod as an example,and the information compression ratio was deduced and calculated by selecting different Angle of the center of the circle.When shooting the condom on the mold rod,three images were collected for each mold rod by shooting one image every 120°.Based on the analysis of cylindrical projection theory,the formula of cylindrical expansion of image was derived,and the algorithm was designed according to the actual situation.Thirdly,the defect detection and marking on the surface of the condom was designed.The algorithm of image enhancement processing was studied,and the most suitable one was selected according to the experimental situation of different algorithms.The adaptive median filter was used to replace the gaussian filter in the traditional Canny operator,and the gradient direction was increased by 45° and 135°.The improved Canny operator was used to extractthe damaged and adhesive defect edges in the image,and the idea of Hough transform was used to identify adhesive defects.Finally,the client application was written and the defect detection algorithm was verified experimentally.The results show that the algorithm can detect the defects on the condom surface accurately.