• Journal of Aerospace System Engineering
• /
• v.9 no.4
• /
• pp.23-30
• /
• 2015

The focal plane image stabilization for a camera is one of the most effective method that can increases the digital camera's image quality by compensating the vibration disturbance. The optical image stabilization can be implemented by making the focal plane to trace the path of incident light. To control the position of focal plane motion compensating stage precisely, a nonlinear control algorithm has been applied by considering coulomb friction which is nonlinear behavior of the compensator system. In our study, we have analyzed the hand shaking vibration using the gyro sensor, and made a mathematical model of compensating stage containing optical sensor and piezo-actuator. Then the nonlinear control algorithm has been designed and its performance has been verified by experiment. In this study, a friction driven peizo-electric actuator with $1{\mu}m$ resolution and 10mm/s speed has been used for stage movement.

• Journal of Broadcast Engineering
• /
• v.25 no.3
• /
• pp.428-438
• /
• 2020

The jitter in the digital videos lowers the visibility and degrades the efficiency of image processing and image compressing. In this paper, we propose a video stabilizer architecture based on triplet CNN and a method of synthesizing training datasets based on video synthesis. Compared with a conventional deep-learning video stabilization method, the proposed video stabilizer can reduce wobbling distortion.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2015.10a
• /
• pp.101-104
• /
• 2015

Motion vector is that comparing a frame between previous frame and current one about how much moved. Using this motion vector, if move the image object of current frame to former frame, it could be corrected to shake from hand and camera shaking. On this thesis, compared efficiency of block matching using SAD(Sum of Absolute Difference) equation as picking out the motion vector, matching using phase correlation, matching using feature point, block matching using bitplane.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.36 no.6
• /
• pp.645-651
• /
• 2012

In this paper, we propose a new method for stabilizing the image captured from a camera mounted on a buoy robot. In this study, in order to solve the problem of cumulative errors and noise produced by a general gyro sensor measuring the orientation angle of the buoy robot, we propose new method for stabilizing the image. In this method, image processing techniques are combined with a newly designed target mounting mechanism that adapts to wave fluctuations. New target extraction and angle estimation techniques are introduced, along with the new mounting mechanism used for the camera and the target, which produce a stabilized image even if the buoy robot is on fluctuating waves.

• Transactions of the Society of Information Storage Systems
• /
• v.7 no.2
• /
• pp.70-74
• /
• 2011

The mobile camera module is a device to be inserted in the digital device for camera feature. The mobile camera module is being shaken by vibrations such as handshake during the exposure time. The clarity is compromised by these vibrations, thus the vibration is considered as an external disturbance. Moreover the use of mobile camera module has been being expanded for automotive vibration should be considered. These external disturbances can cause image blurring, thus optical image stabilization should be applied for image compensation. The compensator is fulfilled mechanically by movable lens group or image sensor that adjusts the optical path to the camera movement. Open loop control is useful for well-defined systems like compliant mechanism. Notch filter and lead compensator are designed and applied to improve the stability and bandwidth. The final level of image compensating is confirmed by image processing with MATLAB and CODE V to verify the better performance.

• Journal of Intelligence and Information Systems
• /
• v.29 no.2
• /
• pp.115-127
• /
• 2023

Video stabilization is one of the camera technologies that the importance is gradually increasing as the personal media market has recently become huge. For deep learning-based video stabilization, existing methods collect pairs of video datas before and after stabilization, but it takes a lot of time and effort to create synchronized datas. Recently, to solve this problem, unsupervised learning method using only unstable video data has been proposed. In this paper, we propose a network structure that learns the stabilized trajectory only with the unstable video image without the pair of unstable and stable video pair using the Convolutional Auto Encoder structure, one of the unsupervised learning methods. Optical flow data is used as network input and output, and optical flow data was mapped into grid units to simplify the network and minimize noise. In addition, to generate a stabilized trajectory with an unsupervised learning method, we define the loss function that smoothing the input optical flow data. And through comparison of the results, we confirmed that the network is learned as intended by the loss function.

• Journal of Digital Contents Society
• /
• v.19 no.8
• /
• pp.1603-1615
• /
• 2018

In order to achieve the low complexity and area, power in the design of Optical Image Stabilization (OIS) suitable for the smart phone, this paper presents the following design explorations, such as; optimization of gyroscope sampling rate, simple and accurate gyroscope filters, and reduced operating frequency of motion compensation, optimized bit width in ADC and DAC, evaluation of noise effects due to PWM driving. In experiments of gyroscope sampling frequencies, it is found that error values are unvaried in the frequency above 5KHz. The gyroscope filter is efficiently designed by combining the Fuzzy algorithm, to illustrate the reasonable compensation for the angle and phase errors. Further, in the PWM design, the power consumption of 2MHz driving is shown to decrease up to 50% with respect to the linear driving, and the imaging noises are reduced in the driving frequency above 2MHz driving frequency. The operating frequency could be reduced to 5KHz in controller and 10KHz in driver, respectively, in the motion compensation. For ADC and DAC, the optimized exploration experiments verify the minimum bit width of 11bits in ADC as well as 10bits in DAC without the performance degradation.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2011.07a
• /
• pp.339-340
• /
• 2011

현재의 IT 산업은 개인 컴퓨터에서 벗어나 이동이 가능한 mobile 분야에서 최근 mobile에 smart를 추가한 ubiquitous로 최근 변화하고 있다. 또한 스마트폰의 보급으로 3D 산업은 monitor display 에서부터 mobile display로 이동하고 있는 추세이다. 하지만 현재까지 3D산업의 mobile display는 device의 공급 부족과 콘텐츠 부족으로 인해 실생활에서 사용되고 있지 못하다. 따라서 안드로이드 폰을 기반으로 한 스마트폰을 이용하여 실시간 입체 영상을 구현하여 원활한 실시간 mobile 3D display 의 가능성과 향후 개선 방향을 제안하고자 한다. 본 논문은 Digital Image Stabilization을 사용하여 motion vector를 계산하여 실시간으로 입체 영상을 구현하는 방법을 제안한다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2007.06a
• /
• pp.833-836
• /
• 2007

Nowadays, the rapid increase of the available amount of internet and system performance has revealed urgent need a method of decryption about digital encryption for stabilization of multimedia data transmission. In this paper, we propose a method of decryption of each frame about video data. Also for advanced decryption, we propose color image decryption method through 4-bit binary of $GF(2^m)$ inverse about an each frame.

• Proceedings of the IEEK Conference
• /
• 2000.07a
• /
• pp.433-436
• /
• 2000

In this paper, we present two types of digital image stabilization (DIS) schemes for mobile video communications. In the first scheme, the DIS system, which is used as a preprocessor of the video encoder, compensates the camera’s undesirable shakes before encoding. This method can reduce the bit rate of encoded video sequence by attenuating the prediction error to be encoded. In the second proposed scheme, the DIS system is coupled with the video decoder. The second scheme uses the K-means clustering algorithm to estimate the camera motion using motion vectors decoded from the received video stream. Simulation results show that the first scheme improves coding efficiency, while the second scheme is computationally efficient since it does not require motion estimation.