A New IC Solder Joint Inspection Method for an Automatic Optical Inspection System Based on an Improved Visual Background Extraction Algorithm

Cai, Nian; Lin, Jianfa; Ye, Qian; Wang, Han; Weng, Shaowei; Ling, Bingo Wing‐Kuen

doi:10.1109/tcpmt.2015.2501284

Cited by 48 publications

(27 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The locally statistical modeling inspects the outlier pixels as potential defect pixels, which indicates the images of inspected samples should have similar appearances. Here, we use the improved ViBe algorithm (Cai et al, 2015) to statistically establish six templates, as it has the advantages of high speed and robustness compared with the other statistical modeling methods (Barnich and Van Droogenbroeck, 2009;Xu et al, 2016). The improved ViBe algorithm uses the hue channel of the IC solder joint image to establish the locally statistical model, as the hue channel contains the major information for colors perception.…”

Section: Locally Statistical Modelingmentioning

confidence: 99%

“…Nowadays, many AOI systems have been successfully used in real assembling lines to inspect component placement, electronic components (Kuo et al, 2019) and solder paste printing (Jan oczki et al, 2013) at the end of the SMT process. However, only a few studies have focused on inspection of solder joints especially IC solder joints owing to the challenges of their small sizes and little distinct appearance differences between qualified and unqualified ones (Cai et al, 2015;Wu and Zhang, 2014).…”

Section: Introductionmentioning

confidence: 99%

“…Only one fixed template cannot well characterize the appearances of IC solder joints although the GMM-based method tackles the problems of the SAM. Cai et al (2015) modified visual background extraction (ViBe) algorithm (Barnich and Van Droogenbroeck, 2009) to establish several templates for well locally modeling IC solder joints. Possible defect pixels of the samples were inspected by comparing the samples with the established templates pixel by pixel.…”

Section: Introductionmentioning

confidence: 99%

“…It means that the local characteristics of the samples can be well represented by these modeling strategies to grasp tiny appearance differences in the qualified sample sequence, which is probably sensitive to the noise and the shift inside the architecture of the IC solder joint. To address this sensitivity, a similar discrimination criterion is proposed for most of these methods based on a defect description metric named defect degree or defect score (Cai et al, 2015(Cai et al, , 2016(Cai et al, , 2017. The defect description metric is defined by weighted summarization of the defect pixels.…”

Section: Introductionmentioning

confidence: 99%

“…To deal with the problems of dependency on the locally statistical model and uncertainty of the discrimination inherently existed in the above statistical-modeling-based methods, a localto-global ensemble leaning mechanism is here proposed to evaluate the qualities of IC solder joints. First, qualified IC solder joints are trained to establish a locally statistical model by means of an improved ViBe algorithm in Cai et al, 2015, as it can fast and accurately establish several statistical templates compared with the other modeling methods that usually model only one template. A potential defect image (PDI) can be achieved by comparing the IC solder joint with the trained model.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Automatic optical inspection system for IC solder joint based on local-to-global ensemble learning

Chen

Cai

Wang

et al. 2020

SSMT

View full text Add to dashboard Cite

Purpose Automatic optical inspection (AOI) systems have been widely used in many fields to evaluate the qualities of products at the end of the production line. The purpose of this paper is to propose a local-to-global ensemble learning method for the AOI system based on to inspect integrated circuit (IC) solder joints defects. Design/methodology/approach In the proposed method, the locally statistically modeling stage and the globally ensemble learning stage are involved to tackle the inspection problem. At the former stage, the improved visual background extraction–based algorithm is used for locally statistically modeling to grasp tiny appearance differences between the IC solder joints to achieve potential defect images for the subsequent stage. At the latter stage, mean unqualified probability is introduced based on a novel ensemble learning, in which an adaptive weighted strategy is proposed for revealing different contributions of the base classifier to the inspection performance. Findings Experimental results demonstrate that the proposed method achieves better inspection performance with an acceptable inspection time compared with some state-of-the-art methods. Originality/value The approach is a promising method for IC solder joint inspection, which can simultaneously grasp the local characteristics of IC solder joints and reveal inherently global relationships between IC solder joints.