• Journal of the Korean Society for Precision Engineering
• /
• v.17 no.10
• /
• pp.63-68
• /
• 2000

We present an automatic inspection system which been developed to evaluate the geometric tolerances of the optical fiber connectors with an dimensional accuracy of$\pm0.1\mnm$. The main part of the inspection system comprises a series of machine vision and laser scanning probes to measure the internal and external circle diameters along with concentricity by making the most of advanced edge detection algorithms. Actual experimental results obtained through various repeatability tests demonstrate that the system well satisfies the required industrial demands for in-situ inspection of optical fiber connections in real manufacturing environment.

• International Journal of Precision Engineering and Manufacturing
• /
• v.1 no.1
• /
• pp.56-61
• /
• 2000

We present a fast inspection method of machine vision for in-line quality assurance of liquid crystal displays(LCD) and plasma display panels(PDP). The method incorporates an optical spatial filter in the Fourier plane of the imaging optics to block the normal periodic pattern, extracting only defects real time without relying on intensive software image process. Special emphasis is on designing a collimated white light source to provide high degree of spatial coherence for effective real time Fourier transform. At the same time, a low level of temporal coherence is attained to improve defect detection capabilities by avoiding undesirable coherent noises. Experimental results show that the proposed inspection method offers a detection accuracy of 15% tolerance, which is sufficient for industrial applications.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.219-219
• /
• 2000

A Plate type pipe is used for heat exchange in radiator of a vehicle. The pipe has several rooms through which water flows and heat is dissipated into outside . In the case that there are small holes, cracks or some scratches on the plate, the radiators lost their functions due to Leakage. This may result in overheating of engine in a car. Thus, we need to perform the entire inspection of the plate type pipe in advance before assembling the radiator. In manufacturing process of the plate type pipe, the productive speed is very high and that may be performed continuously. So, there is no room to perform the outlook inspection by typical image processing techniques. This paper proposes a new method to inspect the outlook surface of the plate type pipe automatically with high speed. Especially, the sequential processing technique of an algorithm which detects defects on the surfaces of the plate type pipe is proposed for line scan camera which captures line image. To evaluate the inspection performance, a series of experiments is performed.

• Journal of Institute of Control, Robotics and Systems
• /
• v.22 no.11
• /
• pp.932-936
• /
• 2016

In this paper, we present a method for the inspection of diamond wheels. In total, six items, including height, radius, and angle, need to be checked during the manufacturing of a diamond wheel. Automatic inspection through image processing is presented in this paper. First, a contour corresponding to the boundary of the diamond wheel is extracted from an image. Next, control points are selected by processing the contour. Seven control points are detected and used for the computation of the required item. Detailed procedures for the computation of the height, radius, and angle using control points are presented in this paper. Experimental results show the feasibility of the presented method.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.12
• /
• pp.903-909
• /
• 2018

In this paper, performance analysis of machine vision techniques is presented to improve the convenience and speed of automatic inspection in the industrial field when machine vision is applied to the image not taken in the stationary state, but in the moving state on a conveyer. When the length of cylindrical rods used for automobiles was measured using the edge detection method, the conveying speed increased, and the uncertainty of the boundary between the background and the part image increased, which resulted in a shorter image of the object taken. This paper proposes a linear compensation method to predict the biased errors of the length measurements after examining the pattern of biased and random errors, respectively, with 6 different types of specimens and 7 velocity stages. The length measurement corrected by the linear compensation method had the same accuracy as the stationary state within the speed range of 30 cm/s and could enhance the application capability in automatic inspections.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.15 no.11
• /
• pp.6863-6869
• /
• 2014

A SMT machine is used to place electronic components on a PCB precisely. To place it precisely, after a component inspection, it finds an offset value using a vision camera, and places the precise position on the PCB. In general, to inspect the components with a vision camera, the components stop in front of the camera for inspection, then move to the placement position. On the other hand, if they do not stop in front of a camera, the inspection time will be shortened and the productivity would be increased. In this thesis, when inspecting without stopping in front of a camera, the fastest way among various routes is described. For the gantry passing over a vision camera, both the distance and speed of a gantry moving trajectory were studied, and there was approximately 5% speed increment when using the method suggested in this thesis.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.17 no.2
• /
• pp.77-83
• /
• 2022

In this paper, the two-stage object detection approach is proposed to implement a deep learning-based label inspection system on edge computing environments. Since the label printed on the products during the production process contains important information related to the product, it is significantly to check the label information is correct. The proposed system uses the lightweight deep learning model that able to employ in the low-performance edge computing devices, and the two-stage object detection approach is applied to compensate for the low accuracy relatively. The proposed Two-Stage object detection approach consists of two object detection networks, Label Area Detection Network and Character Detection Network. Label Area Detection Network finds the label area in the product image, and Character Detection Network detects the words in the label area. Using this approach, we can detect characters precise even with a lightweight deep learning models. The SF-YOLO model applied in the proposed system is the YOLO-based lightweight object detection network designed for edge computing devices. This model showed up to 2 times faster processing time and a considerable improvement in accuracy, compared to other YOLO-based lightweight models such as YOLOv3-tiny and YOLOv4-tiny. Also since the amount of computation is low, it can be easily applied in edge computing environments.

• Journal of Korea Society of Industrial Information Systems
• /
• v.27 no.2
• /
• pp.71-82
• /
• 2022

In order to replace the sampling tensile test of products produced in the tab welding process, which is one of the automotive battery manufacturing processes, vision inspectors are currently being developed and used. However, the vision inspection has the problem of inspection position error and the cost of improving it. In order to solve these problems, there are recent cases of applying deep learning technology. As one such case, this paper tries to examine the usefulness of applying Faster R-CNN, one of the deep learning technologies, to existing product inspection. The images acquired through the existing vision inspection machine are used as training data and trained using the Faster R-CNN ResNet101 V1 1024x1024 model. The results of the conventional vision test and Faster R-CNN test are compared and analyzed based on the test standards of 0% non-detection and 10% over-detection. The non-detection rate is 34.5% in the conventional vision test and 0% in the Faster R-CNN test. The over-detection rate is 100% in the conventional vision test and 6.9% in Faster R-CNN. From these results, it is confirmed that deep learning technology is very useful for detecting welding error of lead tabs in automobile batteries.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.16 no.2
• /
• pp.79-79
• /
• 1991

This paper presents an automated visual inspection/measuring system for the door inner panel manufacturing process. The main objective of this research is to develop inspection applications with a machine vision system related to the determination or quantification of parameters such as dimension, shape, quality of surfaces, number of holes in a panel, and presence or absence of specific features. For quality measurements, we showed a subpixel image processing technology that will improve the consistency of results and level of precision. Also presented is a data analysis method th detect process shifts so that corrective action can be undertaken before more defective units are manufactured.

• Journal of the Korean Society of Manufacturing Technology Engineers
• /
• v.19 no.6
• /
• pp.768-775
• /
• 2010

Automotive industry is a major business area in this country and it becomes more and more important. In order to maintain high quality of vehicles, every parts should be inspected. Among them the inspection job of the knuckle bracket holes of the outer tube of shock-absorber has been done manually until now. So, it takes long time and every product can not be inspected. An automatized inspection system was proposed utilizing machine vision technology, which was composed of a slit beam laser, CCD camera, image processing computer, special jig and illuminating back lights. An algorithm which could process images of the laser and bracket holes, then gave the position, radius, roundness of the holes, was developed. This system was applied for the good and no good products and the performance was confirmed.