06-05-2017, 09:31 AM
A new implementation that separates two of the existing clusters containing two defects each into four new clusters each containing a defect by processing synthetic images of single layer PCBs through the bore hole using MATLAB Image Processing. This will also use Mathematical Morphological Toolbox here. The morphological process involves techniques such as dilation, erosion, opening and closing which helps to divide the images and associates certain types of defects with certain patterns. This project separates defects into larger groups into smaller groups. This increases the efficiency of the inspection system by classifying defects. Because certain PCB patterns occur in different processes, defect classification can help determine the root causes of the error and reduce long-run production costs. The approach is given to separate groups of defects in the hole segment, thick line segment and thin line segment.
One of the key components in the electronics industry is the production of printed circuit board (PCB). Since the existing technology is intended for full digital implementation, it is expected that PCB manufacturing will grow more and more. Along with this development, Malaysia has taken an important step. There are currently 37 PCB manufacturing companies across the country. As important as producing the PCB is to produce a zero defect PCB. This is to ensure a high quality PCB which translates into reliable and quality digital end products.
Initially, bare PCBs (PCBs without components attached to it) were randomly inspected using the manual inspection system, which involves human operators. This technique is quite expensive as it is highly prone to errors due to human errors. A more sophisticated way of doing the inspection is the use of internal circuit testing (ICT) technique. This technique uses a very expensive machine that checks the conductivity of the PCB using probes. However, the limitation of this technique is that it can only detect defects that are based on short or open. Many levels and different techniques of image processing can be used.
There have been several concentrated works in the detection of defects in printed circuit boards (PCBs), but it is also crucial to classify these defects to analyze and identify the causes of defects. This project is aimed at detecting and classifying defects in single-layer PCBs discovered by introducing a hybrid algorithm combining research conducted by Heriansyah et al and Khalid . This project proposes a system for detection and classification of PCB defects using an algorithm for segmentation of morphological images and simple theories of image processing . Based on initial studies, some PCB defects can only exist in certain groups. Therefore, it is obvious that the image processing algorithm could be improved by applying a segmentation exercise. This project uses templates and image tests of single layer, bare and grayscale PCBs. The research improves the work of Khalid by increasing the number of defect categories from 5 to 7, each category ranking a minimum of 1 to a maximum of 4 different types of defects and a total of 13 out of 14 defects.
One of the key components in the electronics industry is the production of printed circuit board (PCB). Since the existing technology is intended for full digital implementation, it is expected that PCB manufacturing will grow more and more. Along with this development, Malaysia has taken an important step. There are currently 37 PCB manufacturing companies across the country. As important as producing the PCB is to produce a zero defect PCB. This is to ensure a high quality PCB which translates into reliable and quality digital end products.
Initially, bare PCBs (PCBs without components attached to it) were randomly inspected using the manual inspection system, which involves human operators. This technique is quite expensive as it is highly prone to errors due to human errors. A more sophisticated way of doing the inspection is the use of internal circuit testing (ICT) technique. This technique uses a very expensive machine that checks the conductivity of the PCB using probes. However, the limitation of this technique is that it can only detect defects that are based on short or open. Many levels and different techniques of image processing can be used.
There have been several concentrated works in the detection of defects in printed circuit boards (PCBs), but it is also crucial to classify these defects to analyze and identify the causes of defects. This project is aimed at detecting and classifying defects in single-layer PCBs discovered by introducing a hybrid algorithm combining research conducted by Heriansyah et al and Khalid . This project proposes a system for detection and classification of PCB defects using an algorithm for segmentation of morphological images and simple theories of image processing . Based on initial studies, some PCB defects can only exist in certain groups. Therefore, it is obvious that the image processing algorithm could be improved by applying a segmentation exercise. This project uses templates and image tests of single layer, bare and grayscale PCBs. The research improves the work of Khalid by increasing the number of defect categories from 5 to 7, each category ranking a minimum of 1 to a maximum of 4 different types of defects and a total of 13 out of 14 defects.