Detailed Research

Product development using various sensors


Fan Unit Motor Condition Monitoring with IMU and PZT Accleration Sensors

2022. 12

Sunha Lee, Jongwook Si, Sungmin Hong, Sungyoung Kim, and Hyungman Lee

Download


The IMU sensor and the PZT acceleration sensor were applied to compare the vibration data of the motor. The sampling rates of the two sensors were set to 1000 Hz, and the acceleration sensor obtained three-axis acceleration data over time using a data collection device (DAQ), and the IMU sensor also obtained three-axis acceleration data over time. Acceleration data according to frequency was obtained through FFT analysis. In addition, research was conducted to develop an artificial intelligence abnormality diagnosis analysis algorithm and learning model using the obtained data.

Development of a Touch Display for Automatic Control of Air Purifier

2022. 06

Moonnyeon Kim, Jongwook Si, and Sungyoung Kim

Download


This system detects surrounding fine dust and efficiently uses the power of the air purifier to maximize power efficiency. In addition, the power of the air purifier can be efficiently controlled by utilizing the concentration of fine dust detected indoors and the concentration of fine dust identified using public data.

Development of a Train Reservation Application Voice Guide System Using RGB Camera

2021. 06

Jongwook Si, Changseop Lee and SungYoung Kim

Download


In this paper, we show a system to help book mobile tickets based on RGB cameras. The mobile phone is detected first then the page is recognized and the user is notified by voice according to the specified scenario. It then shows a system that detects fingertips and recognizes contact using text areas to help with payment steps.

Pan-Tilt Angle Mapping Method using Monocular Camera and Point-to-Point LiDAR Sensor

2021. 06

Hyeoncheol Son, Moonnyeon Kim, Jongwook Si, Changseop Lee and Sungyoung Kim

Download


In this paper, we propose to map the coordinates of images of point-to-point LiDAR sensors and cameras by manipulating the pan-tilt. Point-to-point LiDAR sensors are used to obtain the desired coordinate distance information through pan-tilt motion control without knowing the camera's viewing angle. By both conditions, the movement from the origin of the pan-tilt and the pixel position from the center of the camera have approximations that are almost the same. The two conditions are symmetrical to each other, and experimental simple approximate values can be obtained. Approximate value can be obtained by using the pan-tilt angle between the camera’s horizontal end position and the object in the center position, and the pan-tilt angle between the camera’s vertical end position and the object in the center position.

Golf Swing Analysis Using Multiple Golf Swing Trajectories

2022. 03

Jongwook Si and Sungyoung Kim

Download


This paper proposes a method of analyzing and evaluating golf swings using Kinect, focusing on the indoor golf industry. The hotelling transformation matrix is obtained using several swings in the conparative group, and the original group swing is conceived in this matrix to generate a converted trajectory. In addition, the similarity between the swing of the existing original group and the swing of the converted original group is converted into a score and presented. It was proved that it was a valid method through experiments according to the number of eigenvectors and the number of controls.

Golf Swing Trajectories Comparison Based on Principal Component Analysis : A Preliminary Study

2021. 06

Jongwook Si and Sungyoung Kim

Download


In this paper, we propose a preliminary study of a system that compares the swings of the original and comparative groups by each interval, comparing them through principal component analysis. For the original group, we take two swings as input and reconstruct the trajectory using the largest eigenvalue through principal component analysis, which compares the trajectory with the control group to present the final score using correlation.

Implementation of Golf Swing Analysis System Based on Swing Trajectories Analysis

2019. 12

Hohan Kim and Sungyoung Kim

Download


In this paper, we describe a new swing analysis system. This system can calculate the similarity between the two trajectories obtained from Kinect to determine the similarity of swing trajectories of different golfers. Input trajectories are resampled to have equal spacing and are performed scaling and translation for accurate trajectory comparison. We have verified the usefulness of the proposed system through various analyzes.

Development of Technique on Correcting Golf Posture Based on Posture Comparison

2019. 06

Hyeoncheol Son, Hohan Kim and Sungyoung Kim

Download


We provide a program with six views of the position change graph for a specific skeleton point, 2D view, data view, and correction view, and with comparing the motion between the original group (own golf action) and the control group (professional golfer golf action). The method proposed in this paper uses skeleton information and depth information for motion captured in real time through Kinect and divides the golf posture into four sections and uses it for comparison.In addition, it is possible to generate and use skeleton information from existing video images using Openpose.

Analysis of Golf Swing by Using Kinect

2018. 06

Hanbyeol Kim, Soohyeong Jang, HoKyun Yoo, JunWoo Lee, YunHye Jeon, DaSul Kim and Sungyoung Kim

Download


In this paper we proposed a method on a golf swing classification by using Kinect v2, a ToF camera. The proposed method automatically classifies four key aspects of the golf swing based on the 2D human joint coordinate information obtained from Kinect. We confirmed the performance of about 30 swing data from 7 golfers. The proposed method can classify golf swing in a relatively inexpensive environment and can be used to develop a self - swing analysis system.