Electric contacts inspection using machine vision
Introduction
In the electrical industry, electric contacts (EC) are important electrical components that connect or disconnect electric circuits between electrical devices. They are widely used in switches, breakers, magnetic contactors and relays, which are commonly used key electronic pieces in motors, computers, home appliances, aerospace, and other industries. For those products adopting switches, breakers, magnetic contacts and relays as components, the quality of electric contacts plays a critical role for their performance and reliability. Every electric contact must be stable and reliable for million of times it makes contact during its life time. Therefore, a variety of analytic and inspecting processes have been developed to ensure their quality during the steps of material incoming, processing and final goods. Electric contacts are generally classified into three types: rivets, buttons, and tips. They are mainly composed in two parts, head and shank, and can be formed into various shapes as shown in Fig. 1. In order to speed up the delivery, to reduce manufacturing cost and to improve production capacity, the standardization of shapes and specification of electric contacts has been an important task. However, due to different specification requirements such as head types, head diameters and thickness, shank diameters and lengths, electric contact with defects such as cracks, burrs, eccentricity of contacts (due to asymmetry when bonding head and shank) and rough surface, may cause infeasible electric conductivity and heat conduction of contacts or block the auto-feeding and aiming-at during the auto-bonding process. Currently, the inspection for appearance of electrical contacts in manufacturing processes depends mainly on skilled human inspectors. Unfortunately, those human inspectors generally have problems such as subjectivity in gauging quality variably, and accuracy. Therefore, automatically inspecting those tiny metallic surface defects becomes an important issue.
Machine vision has been a widely used technology in industry for the past three decades. It has been an excellent tool for many industrial inspection tasks such as textiles [20], printed circuit boards [30], electric components [11], chip alignment, wire bonding for Integrated Circuit (IC) [12], [27], machine tools [25] and profile matching for machined parts [7]. Various techniques used for these inspection applications have been reviewed by Chin [6], Newman and Jain [19], Thomas et al. [26] and Malamas et al. [14]. In order to achieve high efficiency and effectiveness in production, the development of automated visual inspection systems is an essential task. In particular, detecting surface defects of metal components is one of the most difficult and commonly seen problems. A variety of computer vision inspection techniques of metal surface defects have previously been discussed in literature. Suresh et al. [23] inspected hot steel slabs with an automated visual inspection system, which compiled statistics for the ratios of length to area of scratch candidates to identify defects. Batchelor and Cotter [1] applied some general image processes like dilation, erosion, and thinning to detect spots and streaks of pollution. Don et al. [8] applied image processing techniques to metal surface inspection, in which both feature extraction and pattern classification were proposed. Xian et al. [31] presented three methods: (1) averaged based shading correction, (2) background model for pollution erasing, and (3) a local iterative operator for defect and background clustering on the surface of bearing rollers. Wong et al. [28] employed fuzzy theory to identify the ratios of length to area of scratch candidates for casting surface under a machine vision system. Nadabar et al. [18] developed a machine vision system to inspect the surface and profile defects for the discrete stamped metal parts with fast inspecting. McBride et al. [17] proposed a vision-based method that enhanced the three-dimensional surface profile to evaluate the arc erosion on electrical contract. Meriaudeau et al. [16] proposed a prototype, using various lighting systems, to provide a view of the whole surface of the tube to the operator so as to reveal the maximum number of defects such as tool marks, scratches and so forth. Martinez-Antón et al. [15] developed a prototype for on-line detection of surface defects in metallic wires, especially for scratches. Zheng et al. [32] proposed an automatic inspection method of metallic surface defects, in which they employed genetic algorithms to automatically learn morphology processing parameters. Wu and Hou [29] proposed an automated visual inspection method for inspecting metal surfaces. They used the modified grey-level co-occurrence matrices of metal images to access the information of metal surfaces. Lin [13] proposed a wavelet characteristic based approach for the automated visual inspection of ripple defects in the surface barrier layer (SBL) chips of ceramic capacitors. Blasco et al. [2] used computer vision to detect peel defects in citrus by means of a region oriented segmentation algorithm. Steiner and Katz [22] used computer vision to inspect porous flaws on machined surfaces. However, the issue of EC surface defect inspection has not been discussed and explored previously. Many other surface defect inspection techniques such as ultrasonic, laser, magnetic, ultraviolet, and X-ray have been developed [17], but those techniques are either costly, slow or have some limitations in environment or setup.
Therefore, the objective of this study is to develop a machine vision-based system to detect the surface defects of electric contact such that the dependability on human inspectors can be relieved and the quality of EC can be improved. The remainder of this paper is organized as follows. In Section 2, the proposed inspection procedures of three different views are presented. Section 3 describes the implementation of the proposed method and conclusions are drawn in Section 4.
Section snippets
Proposed method
This study proposes machine vision-based inspection system to inspect the three views of electric contacts. Their acquired image and corresponding dimensions are shown in Fig. 2. The surface defects discussed in this study include: deckle edge, extra-metal, edge break from the top view; back cracks, and eccentricity from the bottom view; and side cracks from the in side view, which are the major defects of electric contact as illustrated in Fig. 3. In particular, the difficulty of inspecting EC
Implementation
The proposed inspection system was implemented on a personal computer with Pentium IV-3.2G CPU with 512 M RAM. A 640 × 480 grey image of electric contact was digitized using the Eurasys Piccolo pro-II frame grabber in a dark background. For the lighting device, a red circular light emitting diode (LED) that provided the best illumination when acquiring metallic contacts was chosen, and its structure of illumination of top, back and side views is illustrated in Fig. 12a–c. The proposed machine
Conclusions
This study develops a machine vision-based inspection system that inspected six types of EC defects, including deckle edge, extra metal, edge breaks, eccentricity, back cracks, and side cracks. The top, side, and bottom views of an EC were first digitized and then a straightforward inspection procedure, including some commonly used image processing techniques such as threshold, morphological operations, median filter, and blob analysis, was conducted. Using 229 samples of ECs, this study
References (32)
- et al.
Detection of cracks using image processing algorithms implemented in hardware
Image and Vision Computing
(1983) - et al.
Computer vision detection of peel defects in citrus by means of a region oriented segmentation algorithm
Journal of Food Engineering
(2007) Roundness measurements for discontinuous perimeters via machine visions
Computer in Industry
(2002)- et al.
A stochastic optimization approach for roundness measurements
Pattern Recognition Letters
(1999) Automated visual inspection: 1981 to 1987
Computer Vision, Graphics, and Image Processing
(1988)- et al.
Machine vision system for inspecting electric plates
Computer in Industry
(2002) Automated visual inspection of ripple defects using wavelet characteristic based multivariate statistical approach
Image and Vision Computing
(2007)- et al.
A survey on industrial vision systems, applications and tools
Image and Vision Computing
(2003) - et al.
A survey of automated visual inspection
Computer Vision and Image Understanding
(1995) - et al.
The automated inspection of lace using machine vision
Mechatronics
(1995)
Linear-time connected-component labeling based on sequential local operation
Computer Vision and Image Understanding
Real-time industrial visual inspection: a review
Real-Time Image
Automated inspection of printed circuit boards through machine vision
Computer in Industry
Automatic inspection of metallic surface defects using genetic algorithms
Journal of Materials Processing Technology
Roundness inspection strategies for machine visions using non-linear programs and genetic algorithms
International Journal of Production Research
Automated visual inspection: a survey
IEEE Transaction on Pattern Analysis and Machine Intelligence, PAMI-4
Cited by (74)
Defect detection of LGP based on combined classifier with dynamic weights
2019, Measurement: Journal of the International Measurement ConfederationA computer vision system for identification of granite-forming minerals based on RGB data and artificial neural networks
2018, Measurement: Journal of the International Measurement ConfederationCitation Excerpt :In this sense, the automatic control of the laser cleaning process requires the rapid and reliable identification of the mineral grains on the surface to be irradiated in order to avoid damages caused by laser overexposure [6,9]. In that direction, computer vision systems are powerful tools for the automatic inspection as it was already demonstrated in many areas of industrial activity; among them, food industry, textile production or electric industry [12–14]. Computer vision techniques refer to images taken with a camera and after processed and analyzed by computer algorithms.
Defect detection on solar cells using mathematical morphology and fuzzy logic techniques
2024, Journal of Optics (India)Investigation of visual inspection methodologies for printed circuit board products
2023, Journal of Optics (India)