Low Cost SMT Inspection System Using Various Image Processing Techniques and Fuzzy Logic

Lew, King Yeow (2018) Low Cost SMT Inspection System Using Various Image Processing Techniques and Fuzzy Logic. Final Year Project (Bachelor), Tunku Abdul Rahman University College.

Text Lew King Yeow.pdf Restricted to Registered users only Download (2MB)

Abstract

Over the last few decades, the market demand of electronic devices continue to increase exponentially. Thanks to the introduction of Surface Mounting Technology in PCB productions, the process of manufacturing electronic products is getting faster and reliable. SMT stands for Surface Mounting Technology was developed in the 1960s and became popular since 1980s. It is a process where components are mounted directly on the surface of PCB board or flex circuits. SMT undergoes more than 50 years of improving, research and development to become a matured PCB manufacturing techniques. The technology is proved to significantly increase the productivity and yield of the traditional through-hole technology where components are inserted into drilled holes of the circuit board. Defects detection are undeniably important to PCB manufacturer to produce high quality electronic products. The traditional visual inspection process was done by human operator under microscope where it is usually very time consuming while the results are unsatisfying. For this reason, many researchers started to came out with new idea to replace the tedious manual visual inspection with machine inspection called “Automated Optical Inspection” (AOI). However, the price for AOI is such a burden that the manufacturer could not afford to implement the system in their process. In this paper, a low cost inspection process to detect mounting defects is proposed. The proposed inspection system is a combination of image processing and fuzzy logic expert system. The main idea of the project is to capture the image by using eyepiece camera mounted on the microscope, and process the image under MATLAB. The proposed system are proved to have a 100% True Negativity (ability to reject PCB with defects) and prevent them from escaping the inspection process. Furthermore, the total cost of this system is only RM240 which is very low compared to AOI machines in the market. It also able to analyze the system in a high processing speed. The system takes 0.318 seconds averagely for each analysis and still able to produce a high accuracy output. At the end of the inspection system, it is able to generate an analysis report and send to the receiver emails.