• The Journal of Korea Robotics Society
• /
• v.14 no.4
• /
• pp.357-362
• /
• 2019

This paper proposes a model and train method that can real-time detect objects and distances estimation based on a monocular camera by applying deep learning. It used YOLOv2 model which is applied to autonomous or robot due to the fast image processing speed. We have changed and learned the loss function so that the YOLOv2 model can detect objects and distances at the same time. The YOLOv2 loss function added a term for learning bounding box values x, y, w, h, and distance values z as 클래스ification losses. In addition, the learning was carried out by multiplying the distance term with parameters for the balance of learning. we trained the model location, recognition by camera and distance data measured by lidar so that we enable the model to estimate distance and objects from a monocular camera, even when the vehicle is going up or down hill. To evaluate the performance of object detection and distance estimation, MAP (Mean Average Precision) and Adjust R square were used and performance was compared with previous research papers. In addition, we compared the original YOLOv2 model FPS (Frame Per Second) for speed measurement with FPS of our model.

• The Journal of Korea Robotics Society
• /
• v.17 no.1
• /
• pp.25-31
• /
• 2022

Moving among multiple floors is one of the most challenging tasks for indoor autonomous robots. Most of the previous researches for indoor mapping and localization have focused on singular floor environment. In this paper, we present an algorithm that creates a multi-floor map using 3D point cloud. We implement localization within the multi-floor map using a LiDAR and an IMU. Our algorithm builds a multi-floor map by constructing a single-floor map using a LOAM-based algorithm, and stacking them through global registration that aligns the common sections in the map of each floor. The localization in the multi-floor map was performed by adding the height information to the NDT (Normal Distribution Transform)-based registration method. The mean error of the multi-floor map showed 0.29 m and 0.43 m errors in the x, and y-axis, respectively. In addition, the mean error of yaw was 1.00°, and the error rate of height was 0.063. The real-world test for localization was performed on the third floor. It showed the mean square error of 0.116 m, and the average differential time of 0.01 sec. This study will be able to help indoor autonomous robots to operate on multiple floors.

• Archives of Plastic Surgery
• /
• v.36 no.6
• /
• pp.691-701
• /
• 2009

Purpose: Deviations of arterial palmar arches in the hand can be explained on the embryological basis. The purpose of this study was to provide new information about palmar arches through cadaver's dissection. The values of the location and diameter in these vessels were analyzed in order to support anatomical research and clinical correlation in the hand. Methods: The present report is based on an analysis of dissections of fifty - three hands carried out in the laboratory of gross anatomy. A reference line was established on the distal wrist crease to serve as the X coordinate and a perpendicular line drawn through the midpoint between middle and ring fingers, which served as the Y coordinate. The coordinates of the x and y values were measured by a digimatic caliper, and statistically analyzed with Student's t - test. Results: Complete superficial palmar archs were seen in 96.2 % of specimens. In the most common type of males, the superficial arch was formed only by the ulnar artery. In the most common type of females, the superficial arch was formed anastomosis between the radial artery and the ulnar artery. The average length of the superficial and deep palmar arch is $110.3{\pm}33.0mm$ and $67.9{\pm}14.0mm$ respectively. Regarding the superficial palmar arch, ulnar artery starts $-16.1{\pm}5.1mm$ on X - line, and $2.5{\pm}24.5mm$ on Y - line. Radial artery appears on palmar side $7.7{\pm}3.2mm$ on X - line, and $20.9{\pm}10.9mm$ on Y - line. But radial artery starts on $6.3{\pm}3.6mm$ on X - line, and $3.4{\pm}5.1mm$ on Y - line. Digital arteries of superficial palmar arch starts on $6.1{\pm}3.7mm$, $33.9{\pm}8.8mm$ on index finger, $1.8{\pm}3.4mm$, $40.1{\pm}7.3mm$ on middle finger, $-3.2{\pm}4.9mm$, $42.6{\pm}7.0mm$ on ring finger, and $-8.9{\pm}5.1mm$, $42.5{\pm}80mm$ on little finger in respective X and Y coordinates. Radial artery of deep palmar arches measured at the palmar side perforating from the dorsum of hand. It's coordinates were $9.7{\pm}4.8mm$ on X - line, $21.7{\pm}10.2mm$ on Y - line. Ulnar artery was measured at hypothenar area, and it's coordinates were $-20.4{\pm}6.3mm$ on X - line, and $30.6{\pm}7.4mm$ on Y - line. Conclusions: Anatomically superficial palmar arch can be divided into a complete and an incomplete type. Each of them can be subdivided into 4 types. The deep palmar arch is less variable than the superficial palmar arch. We believe these values of the study will be used for the vascular surgery of the hand using the endoscope and robot in the future.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2004.10a
• /
• pp.127-130
• /
• 2004

For the quality control of the industrial products, an automatic hole measuring system has been developed. The measurement device allows X-Y movement due to contact forces between a hole and its own circular cone and the device is attached to an industrial robot. Its measurement accuracy is about 0.04mm. This movement of the plate is measured by two LVDT sensor system. But this system using the LVDT sensors is restricted by high cost and precision of measurement and correspondence of environment so particularly, a vision system with CCD-Camera is discussed in this paper for the above mentioned purpose. The device consists of two of two links jointed with hinge pins basically and, they guarantee free movement of the touch prove attached on the second link in the same plane. These links are returned to home position by the spring plungers automatically after each process for the next one. On the surface of the touch prove, it has a circular white mark for camera recognition. The system detect and notify the center coordinate of capture mark image through the image processing. Its measuring accuracy has been proved to be about $\pm$0.01mm through the repeated implementation over 200 times. This technique will shows the advantage of touch-indirect image capture idea using cone-shaped touch prove in various symmetrical shaped holes particulary, like tapped holes, chamfered holes, etc As a result, we attained our object in a view of the accuracy, economical efficiency, and functionality

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.42 no.1
• /
• pp.1-11
• /
• 2005

Several Solid Freeform Fabrication Systems(SFFS) are commercialized in a few companies for rapid prototyping. However, they have many technical problems including the limitation of applicable materials. A new method of agile prototyping is required for the recent manufacturing environments of multi-item and small quantity production. The objectives of this paper include the development of a novel method of SFFS, the CAFL/sup VM/(Computer Aided Fabrication of Lamination for Various Material), and the manufacture of the various material samples for the certification of the proposed system and the creation of new application areas. For these objectives, the technologies for a highly accurate robot path control, the optimization of support structure, CAD modeling, adaptive slicing was implemented. However, there is an important problem with the conventional 2D lamination method. That is the inaccuracy of 3D model surface, which is caused by the stair-type surface generated in virtue of vertical 2D cutting. In this paper, We design the new control algorithm that guarantees the constant speed, precise positioning and tangential cutting on the 5DOF SFFS. We develop the tangential cutting algorithm to be controlled with constant speed and successfully implemented in the 5DOF CAFL/sup VM/ system developed in this paper. Finally, this paper confirms its high-performance through the experimental results from the application into CAFL/sup VM/ system.

• Journal of Bio-Environment Control
• /
• v.13 no.1
• /
• pp.1-7
• /
• 2004

This study is kinematics for the manipulator development of cucumber harvesting. A theory value was verified by repeated error measurement after the forward kinematics or inverse kinematics analysis of manipulator. Manipulator is consisted of one perpendicular link and two revolution link. The transformation of manipulator can be valued by kinematics using Denavit-Hartenberg parameter. The value of inverse kinematics which is solved by three angles faction shows two types. Repeated errors refered maximum 2.60 mm, 2.05mm and 1.55 mm according to X, Y, Z axis. In this study, the actual coordinates of maximum point and minimum point were agreement in the forward kinematics or inverse kinematics. The results of repeated error measurement were reflect to be smaller compared to a diameter of cucumber. measurement errors were determined by experimented errors during the test. For reducing errors of manipulator and improving work efficiency, the number of link should be reduced and breeding and cultural environment should be considered to reduce the weight and use the hard stuff. The velocity of motor for working should be considered, too.

• Journal of Biomedical Engineering Research
• /
• v.41 no.5
• /
• pp.173-178
• /
• 2020

Remote control intervention surgery robotic system improves treatment effect on cardiovascular patients and reduces X-ray exposure. However, at the time of the first procedure, CT (computerized tomography) and other ultrasound diagnostic equipment should be used because the operator must insert the cannula directly into the patient's leg. Improvements to this have been un-met-needs of hospitals. In this paper, we developed a system that can insert the cannula intuitively and quickly by displaying blood vessels at a glance through the system using smart wearable glasses. The core development method is as follows. In order to project augmented reality onto the surgical image, CT scan angiography image is extracted and processed. In the process, three CT-Markers are used to create a coordinate system of blood vessel images. Additionally, a reference marker is photographed on a single camera to obtain a camera coordinate system. Since the CT marker and the reference marker are in the same position, 3D registration is performed. In the text, a detailed explanation will be given.

• The Journal of the Korea Contents Association
• /
• v.7 no.3
• /
• pp.1-8
• /
• 2007

The images, which were acquired from various systems such as CCTV and Robot, include many human faces. Because of a rapid increase in visual data, we cannot process these manually; rather we need to do these automatically. Furthermore, companies require automatic security systems to protect their new technology. There are various options available to us, including face recognition, iris recognition and fingerprint recognition. Face recognition is preferable since it does not require direct contact. However, the standard 2-Dimensional method is limited, so Morphable Models may be recommended as an alternative. The original morphable model, made by MPI, contains a large quantity of data such as texture and geometry data. This paper presents a Geometrix-based morphable model designed to reduce this data capacity.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.52 no.8
• /
• pp.132-139
• /
• 2015

Visual SLAM(Simultaneous Localization And Mapping) has been used widely to estimate a mobile robot's location. Visual SLAM estimates relative motions with static visual features over image sequence. Because visual SLAM methods assume generally static features in the environment, we cannot obtain precise results in dynamic situation including many moving objects: cars and human beings. This paper presents a stereo vision based SLAM method in dynamic environment. First, we extract disparity map with stereo vision and compute optical flow. We then compute disparity change that is the estimated flow field between stereo views. After examining the disparity change value, we detect ROIs(Region Of Interest) in disparity space to determine dynamic scene objects. In indoor environment, many structural planes like walls may be determined as false dynamic elements. To solve this problem, we segment the scene into planar structure. More specifically, disparity values by the stereo vision are projected to X-Z plane and we employ Hough transform to determine planes. In final step, we remove ROIs nearby the walls and discriminate static scene elements in indoor environment. The experimental results show that the proposed method can obtain stable performance in dynamic environment.