Automated visual inspection in the semiconductor industry: A survey

In this paper, we develop and put into practice an automatic optical inspection (AOI) system based on machine vision to check the holes on a printed circuit board (PCB). We incorporate the hardware and software. For the hardware part, we combine a PC, the three-axis positioning system, a lighting device, and charge-coupled device cameras. For the software part, we utilize image registration, image segmentation, drill numbering, drill contrast, and defect displays to achieve this system. Results indicated that an accuracy of 5 $\mu \text{m}$ could be achieved in errors of the PCB holes allowing comparisons to be made. This is significant in inspecting the missing, the multi-hole, and the incorrect location of the holes. However, previous work only focuses on one or other feature of the holes. Our research is able to assess multiple features: missing holes, incorrectly located holes, and excessive holes. Equally, our results could be displayed as a bar chart and target plot. This has not been achieved before. These displays help users to analyze the causes of errors and immediately correct the problems. In addition, this AOI system is valuable for checking a large number of holes and finding out the defective ones on a PCB. Meanwhile, we apply a 0.1-mm image resolution, which is better than others used in industry. We set a detecting standard based on 2-mm diameter of circles to diagnose the quality of the holes within 10 s.

A Machine Vision Based Automatic Optical Inspection System for Measuring Drilling Quality of Printed Circuit Boards

Electronics industry is one of the fastest evolving, innovative, and most competitive industries. In order to meet the high consumption demands on electronics components, quality standards of the products must be well-maintained. Automatic optical inspection (AOI) is one of the non-destructive techniques used in quality inspection of various products. This technique is considered robust and can replace human inspectors who are subjected to dull and fatigue in performing inspection tasks. A fully automated optical inspection system consists of hardware and software setups. Hardware setup include image sensor and illumination settings and is responsible to acquire the digital image, while the software part implements an inspection algorithm to extract the features of the acquired images and classify them into defected and non-defected based on the user requirements. A sorting mechanism can be used to separate the defective products from the good ones. This article provides a comprehensive review of the various AOI systems used in electronics, micro-electronics, and opto-electronics industries. In this review the defects of the commonly inspected electronic components, such as semiconductor wafers, flat panel displays, printed circuit boards and light emitting diodes, are first explained. Hardware setups used in acquiring images are then discussed in terms of the camera and lighting source selection and configuration. The inspection algorithms used for detecting the defects in the electronic components are discussed in terms of the preprocessing, feature extraction and classification tools used for this purpose. Recent articles that used deep learning algorithms are also reviewed. The article concludes by highlighting the current trends and possible future research directions.

/pdf/a-review-and-analysis-of-automatic-optical-inspection-and-3pj4dl25sg.pdf

A Review and Analysis of Automatic Optical Inspection and Quality Monitoring Methods in Electronics Industry

The paper considers the problems of scheduling n jobs that are released over time on a machine in order to optimize one or more objectives. The problems are dynamic single-machine scheduling problems (DSMSPs) with job release dates and needed to be solved urgently because they exist widely in practical production environment. Gene expression programming-based scheduling rules constructor (GEPSRC) was proposed to construct effective scheduling rules (SRs) for DSMSPs with job release dates automatically. In GEPSRC, Gene Expression Programming (GEP) worked as a heuristic search to search the space of SRs. Many experiments were conducted, and comparisons were made between GEPSRC and some previous methods. The results showed that GEPSRC achieved significant improvement.

Evolving scheduling rules with gene expression programming for dynamic single-machine scheduling problems

Computer-aided visual inspection of surface defects in ceramic capacitor chips

This paper studies automatic optical inspection for detecting defects on the printed circuit board inner layer. The development of this study can be divided into five stages, they are reference image rebuilding, inspection image normalization, image subtraction, defects separation and defect classification. In the image subtraction stage, the difference between the reference image from the printed circuit board design and the inspected image is checked for defects. Each defect region is separated using a defect outer boundary tracing method. A boundary state transition method is proposed to classify the defect types. This system can recognize eight defect types, open, mouse bite, pinhole, missing conductor, short, spur, excess copper and missing hole. In addition, a comparison with the methods described in the literature is made, proving that the proposed method produces better results .

Automatic optical inspection for detecting defects on printed circuit board inner layers

Allocating inspection stations is a very important issue for controlling quality and saving cost. The study considers inspection allocation problems for serial production systems with two types of workstations, workstation of attribute data and workstation of variable data, which better represents real practice. In addition, this study considers three possibilities for the treatment of detected nonconforming units, namely, repair, rework and scrap. With the above considerations, a profit model is developed for optimally allocating inspection stations. Moreover, a heuristic solution method is developed and it is proved to have much less computation time, compared with an optimization method based on complete enumeration, especially as the number of workstations increases. The model is highly extensible and applicable. Therefore, it can serve as a production planning guideline to solve the inspection allocation problem for serial production systems .

Inspection allocation planning with two types of workstation: WVD and WAD

This paper studies dynamic selection of sequencing rules for a class-based unit-load automated storage and retrieval system (AS/RS). A multi-pass and genetic algorithm (MPGA) simulation system is developed and it divides storage and retrieval requests or dual commands into a series of blocks, and then conquers each block to find the most promising combination of sequencing rules. These rules are first come first served (FCFS), shortest total-travel time (STT) and shortest due time (SDT), and they can be chosen dynamically in any decision points in the system. An experiment shows that our approach with dynamic rules is much better than those approaches with any single rule used from the beginning to the end in the whole system. The results of this study provide a better way to control and manage the operation of AS/RS.

Dynamic selection of sequencing rules for a class-based unit-load automated storage and retrieval system

A simulation system with the approach of multi-pass and genetic algorithm (MPGA) is developed for dispatching dual commands by a stacker machine in a unit-load/class-based automated storage and retrieval system (AS/RS). Three single-rules, FCFS/STT, SDT/STT, and STTbk, are run individually by a single-pass simulation and served as the benchmarks to compare with the performances of three kinds of the MPGA approach. All experiments, singles and MPGA, start with the same initial system states and run with the same input data streams. Due to a larger solution space for the optimal search and an iterative searching method, the MPGA approach greatly improves the performances of our AS/RS model, compared with traditional approaches.

H. Rau

Papers

Automatic optical inspection for detecting defects on printed circuit board inner layers

Inspection allocation planning with two types of workstation: WVD and WAD

Dynamic selection of sequencing rules for a class-based unit-load automated storage and retrieval system

Dual commands dispatching of a class-based unit-load automated storage and retrieval system using multi-pass simulation with generic algorithm