• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2007.06a
• /
• pp.601-604
• /
• 2007

In this paper dose a laver surface check using a real time image process. This system does false retrieval of a laver at a laver production line. At first, a laver image was read in real time using a CCD camera. In this paper, we use an area scan CCD camera. Image is converted into a binary code image using a high-speed imaging process board afterwards. A laver feature is extracted by a binary code image. Surface false retrieval is finally executed using a laver feature. In this paper, we use an area feature of a laver image.

• Journal of the Korean Society for Precision Engineering
• /
• v.18 no.9
• /
• pp.179-185
• /
• 2001

In this study, a visual sensor system for weld seam tracking the I-butt weld joints in GMA welding was constructed. The sensor system consists of a CCD camera, a diode laser with a cylindrical lens and a band-pass-filter to overcome the degrading of image due to spatters and arc light. In order to obtain the enhanced image, quantitative relationship between laser intensity and iris number was investigated. Throughout the repeated experiments, the shutter speed was set at 1-milisecond for minimizing the effect of spatters on the image, and therefore most of the spatter trace in the image have been found to be reduced. Region of interest was defined from the entire image and gray level of searched laser line was compared to that of weld line. The differences between these gray levels lead to spot the position of weld joint using central difference method. The results showed that, as long as weld line was within $^\pm$15$^\circ$from the longitudinal straight fine, the system constructed in this study could track the weld line successful1y. Since the processing time reduced to 0.05 sec, it is expected that the developed method could be adopted to high speed welding such as laser welding.

• Journal of Animal Environmental Science
• /
• v.8 no.3
• /
• pp.165-170
• /
• 2002

This study was carried out to measure the 3D-distance of a cow model teat for an application possibility on Robotic Milking System(RMS). A teat recognition algorithm was made to find 3D-distance of the model by using Gonzalrez's theory. Some of the results are as follows. 1 . In the distance measurement experiment on the test board, as the measured length, and the length between the center of image surface and the measured image point became longer, their error values increased. 2. The model teat was installed and measured the error value at the random position. The error value of X and Y coordinates was less than 5㎜, and that of Z coordinates was less than 20㎜. The error value increased as the distance of camera's increased. 3. The equation for distance information acquirement was satisfied with obtaining accurate distance that was necessary for a milking robot to trace teats, A teat recognition algorithm was recognized well four model cow teats. It's processing time was about 1 second. It appeared that a teat recognition algorithm could be used to determine the 3D-distance of the cow teat to develop a RMS.

• Proceedings of the IEEK Conference
• /
• 2006.06a
• /
• pp.933-934
• /
• 2006

High level perceptual tasks such as context understanding, SLAM and object recognition are essential for intelligent robot to provide services for human supports. Those intelligent robots often use camera sensor for vision information, sonar or laser sensor for range information, encoder for angular velocity of wheel and so on. The information is generated at different time intervals by the different H/W devices and S/W algorithms. The generation of high level information requires the specific mixture of low level information. And the information should be represented to be useful for robots to use in their ecological niche. In conventional robot systems, perceptual module requires the resource to use by tightly coupling whenever it is needed. So the resource and information cannot be easily shared and even could be invalid for the delayed information. In this paper, we propose a synchronization system of robot-centered information for context understanding. Our system represents information for the robot capacity and synchronizes the information that is asynchronously generated, where is employed the black-board architecture.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.10a
• /
• pp.538-539
• /
• 2018

This paper summarizes the implementation of lane recognition using OpenCV, one of the open source computer vision libraries. The Linux operating system Rasbian(r18.03.13) was installed on the ARM processor-based Raspberry Pi 3 board, and Raspberry Pi Camera was used for image processing. In order to realize the lane recognition, Canny Edge Detection and Hough Transform algorithm implemented in OpenCV library was used and RANSAC algorithm was used to prevent shaking of vanishing point and to detect only the desired straight line. In addtion, the DC motor and the Servo motor were controlled so that the vehicle would run according to the detected lane.

• Journal of Broadcast Engineering
• /
• v.24 no.6
• /
• pp.1064-1071
• /
• 2019

Multiview stereo matching algorithm is used to reconstruct 3D shape from a set of 2D images. Conventional multiview stereo algorithms have been implemented on high-performance hardware due to the heavy complexity that contains a large number of calculations in each step. However, as the performance of mobile graphics processors has recently increased rapidly, complex computer vision algorithms can now be implemented on mobile devices like a smartphone and an embedded board. In this paper we parallelize an multiview stereo algorithm using OpenCL on mobile GPU and provide various optimization techniques on the embedded hardware with limited resource.

• Korean Journal of Optics and Photonics
• /
• v.16 no.4
• /
• pp.345-353
• /
• 2005

We proposed a new and cost-effective method fer assembling holographic pickup modules without any high resolution vision system. Assembling was accomplished by adjusting photodiode package only, leading to a low cost, holographic pickup module. Focus and tracking error signals were simply determined by comparing spot sizes and by using the 3 beam method, respectively, based on four-sectional holographic optical elements. In experiment, we assembled a hologram module and estimated performance of the proposed method fur a holographic pickup module used in compact disc system.

• Journal of Korea Multimedia Society
• /
• v.14 no.5
• /
• pp.632-641
• /
• 2011

This paper presents a position error recognition system when the wafer is mounted in cleaning equipment among the wafer manufacturing processes. The proposed system is to enhance the performance in cost and reliability by preventing the wafer cleaning system from damaging by alerting it when it is put in correct position. The proposed algorithm is in obtaining a mapping function from camera and physical wafer by designing and manufacturing the radial shape calibrator to reduce the error by using the conventional chess board one. The system is to install in-line process using high reliable and high accurate position recognition. The experimental results show that the performance of the proposed system is better than that of the existing method for detecting errors within tolerance.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.11 no.6
• /
• pp.1136-1140
• /
• 2007

In this paper dose a laver surface check using a real time image process. This system does false retrieval of a laver at a laver production line. At first, a laver image was read in real time using a CCD camera. In this paper, we use an area scan CCD camera. Image is converted into a binary code image using a high-speed imaging process board afterwards. A laver feature is extracted by a binary code image. Surface false retrieval is finally executed using a laver feature. In this paper, we use an area feature of a laver image.

• Journal of Korea Multimedia Society
• /
• v.25 no.1
• /
• pp.39-51
• /
• 2022

Image-based gender classification and age estimation of human are classic problems in computer vision. Most of researches in this field focus just only one task of either gender classification or age estimation and most of the reported methods for each task focus on accuracy performance and are not computationally light. Thus, running both tasks together simultaneously on low cost mobile or embedded systems with limited cpu processing speed and memory capacity are practically prohibited. In this paper, we propose a novel light-weight gender classification and age estimation method based on ensemble multitasking deep learning with light-weight processing neural network architecture, which processes both gender classification and age estimation simultaneously and in real-time even for embedded systems. Through experiments over various well-known datasets, it is shown that the proposed method performs comparably to the state-of-the-art gender classification and/or age estimation methods with respect to accuracy and runs fast enough (average 14fps) on a Jestson Nano embedded board.