Use of image processing and pattern recognition for checking surface quality of IC chip with automated optical inspection (AOI) | |
| Author | Naravane, Chaitanya Madhav |
| Call Number | AIT RSPR no.ISE-00-01 |
| Subject(s) | Image processing Pattern perception |
| Note | A research study submitted in partial fulfillment of the requirement for the degree of Master of Engineering, School of Advanced Technologies |
| Publisher | Asian Institute of Technology |
| Series Statement | Research studies project report ; no. ISE-00-01 |
| Abstract | Computer vision has high importance in semiconductor industry because it is fast, accurate, economical and convenient. Hence it is used from the very beginning of IC manufacturing, wafer fab, till the end of process of Test, Mark and Pack. Following work is an attempt to use acuWin, a image processing and pattern recognition software and C++ to check surface deformity on IC chip. To start with, program is written in acuWin’s language called Script using different approaches. Template matching is used as basis for pattern recognition. Each approach use different objects and objects in combination and are compared for the best result. The best result is the one which takes least time for execution and largest difference of score for good surface and one with deformity. Limited number of chips were used during experiments and also only limited number of readings were carried out. Thus results obtained are essentially under constrained environment. It is observed that use of acuWin’s objects ApxMod and ApxSch in combination gives the best result while various neighbourhood operations have not yielded any advantage but rather their results are poor. This approach is taken as basis for further work of developing C ++ program. Also readings were taken with chips tilted at different angles from zero degree through 360 degrees. It showed scores were highest with angle of zero degree and it reduced as angle was increased to 180 degrees. It again increased when angle of the chip was increased to 360 degrees. Results showed that scores for zero degree and 360 degrees which ideally should be the same, varied in magnitude. This can be attributed to lower repeatability of the system or to error of measurent due to small number of measurement.. C++ program is developed using Borland compiler (version 5.02). AcuWin libraries are called while developing this program. A graphical user interface is created and results are saved in Notepad. It is concluded that template matching is insufficient. Optical Character Verification as a basis of pattern recognition would yield better results. Also this work can be extended for checking dimensional accuracy and alignment of pins of chips. |
| Year | 2000 |
| Corresponding Series Added Entry | Asian Institute of Technology. Research studies project report ; no. ISE-00-01 |
| Type | Research Study Project Report (RSPR) |
| School | School of Advanced Technologies (SAT) |
| Department | Department of Industrial Systems Engineering (DISE) |
| Academic Program/FoS | Industrial Systems Engineering (ISE) |
| Chairperson(s) | Manukid Parnichkun; |
| Examination Committee(s) | Afzulpurkar, Nitin;Voratas Kachivichyanukul; |
| Scholarship Donor(s) | Siemens A. G.; |
| Degree | Research Studies Project Report (M. Eng.) - Asian Institute of Technology, 2000 |