| The Surface Mount Technology solder joint (SMT Solder Joint) of Circuit Modules products (SMT products) has the characteristics not only keeping electrical performance smoothly, but also ensuring the reliability of mechanical connection. The reliability of the solder joints is the core of the SMT products. The fault detection of the soldered joints and the control technology of the assembly quality are the key technology to ensure the quality and the reliability of SMT products. The quality detection of the soldered joints develops toward three-dimensional (3D), high precision and low cost. The more quality information can be obtained from 3D structures, comparing with what being obtained from two-dimensional structures, as a result, the higher accuracy can be got. The researches on 3D quality information extraction of SMT solder joint and other related technologies and the products data and the quality information parameter which are comparable with reasonable form, are very useful to improve the automatic detection ability of SMT solder joint quality, and to guide for improving welding technology.The several main research achievements and innovations in this paper:(1) The method of removing the mixed noise in SMT solder joint image was studied. Under the guidance of "noise decomposition, hierarchical processing, combining linear filtering and nonlinear filtering", an AWPWM denoising scheme is proposed based on the wavelet packet (WP) transform combining Wiener filter and Median filter. The method effectively solves the problem of mixed noise in SMT solder joint image. The good performance of the proposed approach is reported and compared. Results obtained from tests and practices showed that, the proposed approach can preserve the image edge-based information well. And it also could obtain the same result in removing a single noise comparing with other conventional methods.(2) Illumination model of SMT solder joint image was studied. To improve the accuracy of SMT solder joint image 3D reconstruction, an improved illumination model suitable for SMT solder joint 3D reconstruction was proposed. Through analyzing the light reflection of SMT solder joint surface, this model not only improved the diffusion reflectance component, but the influence of specular reflectance component with linear superposition was also considered. The variables of the model are dispersed by using finite difference method and the model was solved by using linear method. Examples used in practice show that the model has the characteristics of reconstruction higher accuracy, simple expression and strong practicality. And when the coefficient of diffusion reflectance was between 0.83 and 0.89, the better result can be obtained, and when the coefficient of diffusion reflectance was 0.86, the coefficient of specular reflectance was 0.14, the best result can be obtained.(3) 3D reconstruction technology of SMT solder joint was studied and a method for SMT solder joint 3D reconstruction based on Shape from shading (SFS) technology was proposed. The method was achieved on the exposed surface of 3D reconstruction of visual-based SMT solder joint. According to the problem of image noise and poor reconstruction precision, AWPWM denoising scheme was used to remove the mixed noise in SMT solder joint image, and the improved illumination model was used to perform SMT solder joint in 3D reconstructing. Examples used in different types and morphology solder joint show the satisfactory result. Comparing with traditional illumination model, the improved illumination model can reduce the influence caused by specular reflectance on the reconstruction results very well, and obtain more ideal result. Aiming at resolving the unfavorable problems in reconstruction result not well, a modified method was proposed to rectify the 3D reconstruction results and at last, the ideal results were achieved. The principle and implementation procedure of the proposed 3D reconstruction method demonstrate the universal applicability in visual and exposed SMT solder joint, such as Chip Components, QFP, LCCC, SOP, and so on.(4) Non-contact extraction technology of 3D quality information extraction of SMT solder joint was studied and a novel approach was proposed by using a single image, adopting SFS methods to reconstruct the 3D shape of solder joint. In order to solve the uncertain problem of 3D form based on SFS method, a sample SMT solder joint was used to get the ration relationship between the actual height and the image pixel value, as a result, the SMT solder joint shape was formed, mean while, some techniques were used to translate matrix of SMT solder joint surface height into entity model automatically. Then solder joint quality information of volume, section area and wetting angles were extracted from the entity model. The novel non-contact extraction approach was used in different specifications chip components solder joint, and the results illustrate that the mean precision of volume was 86.82%.(5)A SMT solder joint 3D quality information extraction system was designed and developed. SMT solder joint 3D quality information non-contact and automatic extraction were realized in the system. The system consists of four modules:SMT Solder Joint Image Denoising Processing, Reconstruction Parameters Obtaining, SMT Solder Joint 3D Surface Reconstruction, SMT Solder Joint 3D quality Information Extraction and generates an executable file. By using the system, SMT solder joint 3D quality information can be extracted automatically and saved in database.3D quality information extraction function in assembly quality online real time detection and control system of SMT solder joint was realized.The researches and achievements of this paper resolved the problem of SMT solder joint image denoise dissatisfactory and its 3D quality information non-contact and automatic extraction. The research meets the engineer and reality demands. The method was the universal applicability in visual and exposed SMT solder joint. The results can provide 3D quality information for SMT solder joint assembly quality online real time detection and control system in quality comparison, analysis and intelligence discrimination of solder joint quality.
|
PDF Full Text Request