Research On Detection Method Of Glue Drawing Defect Of Micro Camera Module

With the popularity of electronic products and camera modules,the demand for improving the quality of micro camera modules is increasing.The painting glue is an important process link in the production of miniature camera modules.The product quality is affected by the area,shape and characteristic distribution of the painting glue.Traditional painting glue surface quality defect detection mainly relies on manual detection,which cannot quantitatively analyze defects,and is highly subjective.The manual detection has low efficiency and poor accuracy rate,which is difficult to meet the needs of industrial high-speed automated production.Finally,the topic of painting glue surface defect detection is selected in this paper.Defect features are extracted from visual images.An adaptive multi threshold segmentation model is established,and a multi feature fusion defect detection algorithm is proposed.Based on monocular vision system,a three-dimensional measurement method of painting glue is proposed to realize the accurate detection of painting glue surface defects and the measurement of three-dimensional surface dimensions.The main research contents of the paper are as follows:1.The hardware scheme of defect detection system is built.The hardware and imaging characteristics of monocular vision system are analyzed.The camera imaging model and coordinate system conversion method are described.Combined with the requirements of working space and detection speed in the production of painting glue process.The hardware of monocular vision system is established,including camera,image acquisition card and light source,which provides hardware basis for the design of painting glue surface defect detection algorithm and three-dimensional measurement.2.The characteristics of the glue image is analyzed,the detection algorithm is designed,and the detection software is developed.Firstly,the feature distribution characteristics of the gluepainted images under the irradiation of three-color annular structured light are analyzed.The circuit board is located and some background information is removed based on the template matching method.To solve the problems of uneven ink painting and uneven surface of circuit board,which lead to large brightness difference of glue painting images in the same batch,and it is difficult to segment by using static threshold.An adaptive multi threshold glue segmentation algorithm is proposed.Secondly,according to the feature analysis of law of painting glue image,the features are divided into global features and local features.The painting glue features are measured and extracted by gray mean,variance,area and center distance.A painting glue surface defect detection algorithm based on multi feature fusion is designed to realize the accurate classification of painting glue surface defects.The MCE value,the accuracy,missed kill rate and over kill rate of the glue defect detection method are used as the evaluation indexes to evaluate the glue segmentation and defect detection algorithm.The experimental results are compared with the existing image segmentation algorithm and glue defect detection algorithm.The experimental results verify the effectiveness and feasibility of the segmentation and defect detection algorithm proposed in this paper.Finally,the software interface is designed based on the requirements of painting glue defect detection process and production process.The secondary development of camera and light source is carried out.The required hardware functions,algorithm processing,result display and other modules are integrated into the software interface.The development of painting glue surface defect detection software is finished.3.A three-dimensional measurement method of painting glue surface based on monocular vision is proposed.Firstly,a point light source radiation model is established.The relationship between the image gray value and the imaging parameters of the vision system are analyzed.Secondly,a single LED light bead is regarded as the radiation of point light source.According to the characteristics of LED ring light in monocular vision system,the coupling law between the parameter variables of vision system and the image gray value are analyzed,the mapping model of image gray value and height under monochromatic light are deduced,the height mapping equation of monochromatic light model is established,and the equation is extended to establish the height mapping equation group under trichromatic ring light.Then,the height information of painting glue surface is obtained by solving the height equations.Finally,a three-dimensional measurement experiment is carried out on the mouth-shaped glue,which is compared with the laser triangulation method.The section of the glue is intercepted for error analysis to verify the proposed Validity of 3D measurement methods.