Earticle

본 연구는 도로 위의 객체를 분류하고 그 위치를 추정하기 위해 카메라와 레이저스캐너를 이용한 이종센서 검지 시스 템의 연구를 다루고 있다. 이러한 도로인프라에서의 검지시스템은 ADAS(Advanced Driver Assist System) 및 (반)자동제어 서비스 등의 새로운 C-ITS 서비스에서 요구되는 객체의 위치 정보를 검지할 수 있다. 본 연구에서는 국외 관련 사례를 살펴보고, 카메라와 레이저스캐너를 이용한 검지시스템의 가능성을 살펴보았다. 그 후 이종센서 처리 알고리즘을 제안하 고, 실 도로환경에서 몇 가지 도로상황 시나리오를 설정하여 시험검증을 실시하였다. 그 결과 이종센서 검지시스템으로 차량, 보행자 및 기타 장애물에 대한 검지 및 위치 추정에 대하여 비교적 용이하게 이용될 수 있음을 확인할 수 있었다. 본 연구의 경우 매우 이상적인 조건에서 실험이 실시되었으며, 조도, 날씨 등의 외부환경 조건의 변화에 따른 알고리즘 의 평가가 필요하다. 이러한 연구는 향후 미래의 C-ITS 환경 하에서 객체 검지 기술로 활용될 수 있을 것으로 기대한다.

This paper proposes the new object detection system with two laser-scanners and a camera for classifying the objects and predicting the location of objects on road street. This detection system could be applied the new C-ITS service such as ADAS (Advanced Driver Assist System) or (semi-)automatic vehicle guidance services using object's types and precise position. This study describes the some examples in other countries and feasibility of object detection system based on a camera and two laser-scanners. This study has developed the heterogenous sensor's fusion method and shows the results of implementation at road environments. As a results, object detection system at roadside infrastructure is a useful method that aims at reliable classification and positioning of road objects, such as a vehicle, a pedestrian, and obstacles in a street. The algorithm of this paper is performed at ideal condition, so it need to implement at various condition such as light brightness and weather condition. This paper should help better object detection and development of new methods at improved C-ITS environment.

최근 3D 디스플레이 시장이 급속히 성장함에 따라 3D 입체영상을 기반으로 한 체감형 콘텐츠역시 시장의 화두로 떠오르며 활발한 연구가 이루어지고 있다. 하지만 관련 기술의 빠른 발전에 비해 보다 근본적인 인간 요소적 한계점에 대한 연구는 미비한 상황이며, 이는 3D 입체영상 시장의 큰 걸림돌이 되고 있다. 본 논문에서는 3D 입체영상 콘텐츠의 제작을 위해 입체영상에 대해 원리와 구조를 파악하고 이를 바탕으로 체감형 3D 입체영상 콘텐츠를 제작해봄으로서 작업자의 경험에 의거한 정성적 3D 입체영상 콘텐츠 제작방식의 문제점을 기술하였다. 그리고 이에 대한 해결책으로 휴먼팩터의 선행연구를 기반으로 한 정량적 3D 입체영상 콘텐츠에 제작 지원도구의 필요성을 제시하였다.

World-wided attention about stereoscopy is being subjected. Vitalization on industrial development of 3D stereoscopy has not yet arrived to local. In this paper, for the vitalization of this field, principles and structures on the stereopsis, and insufficiency and drawback, which occur while games are produced, have been studied. If solutions are suggested, and the problems are solved, the contents of 3D stereoscopy can be vitalized.

Following results about Y,M,C Mask are obtained by the Indirect-Screen Color Separation Method. We make use of experimental system which are in use for the student education. In the Camera-Back Method of reflection copy, the results are summerized as follows.

The influence of acquisition environment on music score images captured by a camera has not yet been seriously examined. All existing Optical Music Recognition (OMR) systems attempt to recognize music score images captured by a scanner under ideal conditions. Therefore, when such systems process images under the influence of distortion, different viewpoints or suboptimal illumination effects, the performance, in terms of recognition accuracy and processing time, is unacceptable for deployment in practice. In this paper, a novel, lightweight but effective approach for dealing with the issues caused by camera based music scores is proposed. Based on the staff line information, musical rules, run length code, and projection, all regions of interest are determined. Templates created from inverse filter are then used to recognize the music symbols. Therefore, all fragmentation and deformation problems, as well as missed recognition, can be overcome using the developed method. The system was evaluated on a dataset consisting of real images captured by a smartphone. The achieved recognition rate and processing time were relatively competitive with state of the art works. In addition, the system was designed to be lightweight compared with the other approaches, which mostly adopted machine learning algorithms, to allow further deployment on portable devices with limited computing resources.

SMT장비 인 겐트리는 전자 부품을 피더에서 흡착하여 기판위에 정확히 실장하는 장비이다. 이 때 부품이 정확 히 흡착되었는지를 검사하기 위해 카메라 앞에서 겐트리는 잠시 멈춘다. 본 논문에서는 이동 시간을 단축하기 위해 카메라 앞에서 겐트리가 정지하지 않고 움직이면서 검사를 할 수 있는 방법 을 제시한다. 이 방법은 여러 이동 경로 중 가장 빠른 시간에 이동할 수 있는 경로를 찾아 기존 방법보다 이동 시간이 단축됨을 보여준다. 제시된 방법(moving-motion)은 기존 방법(stop-motion)보다 겐트리의 이동 시간을 20%나 단축 시킨다.

Gentry, an SMT equipment, is a device that absorbs electronic components from the feeder and mounts them correctly on the PCB. At this time, the gantry stops in front of the camera to check whether the parts are correctly adsorbed. In this paper, we propose a method to shorten the moving time by moving the gantry without stopping in front of the camera. This method shows that the moving time is shorter than that of the conventional method by finding a path that can be moved in the fastest time among the various moving paths. The proposed method (moving-motion) reduces the gantry travel time by 20% compared to the conventional method (stop-motion).

표면실장장비(SMD)의 겐트리는 부품을 피더에서 기판으로 옮기는 역할을 한다. 이 때 겐트리의 이동 시간은 생산 량에 영향을 미친다. 따라서 본 논문에서는 겐트리의 이동 시간을 단축시키기 위해 부품의 흡착에서 실장까지 가장 빠른 경로를 제시한다. 이 경로는 겐트리의 특성상 카메라 앞에서의 속도가 2인 경우이다. 따라서 이때의 궤적 그래프를 시뮬레이 션을 통해 작성하고 이동 시간을 계산하였다. 그 결과 현재 사용 중인 stop-motion방법보다 본 논문에서 제시한 moving-motion방법의 이동 시간이 20%단축된 것을 알 수 있다.

The gantry of surface mount equipment (SMD) is responsible for transferring parts from the feeder to the PCB. At this time, the moving time of the gantry affects the yield. Therefore, in this paper, we propose the fastest path from the suction to the mounting to reduce the gantry travel time. This path is a case where the velocity in front of the camera is 2m/sec due to the nature of the gantry. Therefore, the trajectory graph of this case was created through simulation and the travel time was calculated. As a result, we can see that the moving time of the moving-motion method proposed in this paper is 20% shorter than the current stop-motion method.

미소칩을 PCB상에 실장하는 장비에서 그 칩을 이동해주는 역할은 겐트리가 하고 있다. 그런데 겐트리의 기계적 한계 때문에 생산성을 증가시키는 것은 쉽지 않다. 따라서 본 논문에서는 생산성을 높일 수 있는 방법을 소프트웨어적으로 해결하고자 한다. 이를 위해 겐트리의 이동 시간을 단축함으로써 생산성을 향상시킬 수 있는 방법을 제시한다. 먼저, 현재 사용하고 있는 방법 (stop-motion)에 대해 총 이동 시간을 계산해 보았다. 그리고 본 논문에서 제시하는 이동 시간 단축 방법(moving-motion)에 대해서도 총 이동 시간을 계산해 보았다. 이 방법은 카메라 앞에서 멈추지 않고 부품 검사를 함으로써 이동 시간을 줄이는 것이다. 그 결과 본 논문에서 제시한 방법이 16%의 시간을 단축하였음을 보여주었다. 향후에는 다른 타입에 대해서도 시간 계산 방법을 연구하는 것이다.

Gentry is responsible for moving the fine chip in the device that mounts the chip on the PCB. However, it is not easy to increase productivity because of the mechanical limitations of the gantry. Therefore, in this paper, we try to solve the method to increase the productivity by software. For this purpose, we propose a method to improve the productivity by shortening the movement time of the gantry. First, we calculated the total travel time for the current method(stop-motion). In addition, the total travel time is also calculated for the travel time reduction method presented in this paper. This method reduces the travel time by checking parts without stopping in front of the camera. As a result, we showed that the proposed method shortened the time of 16%. In the future, we will study time calculation methods for other types.

SMT장비는 전자 부품을 흡착하여 PCB상에 정확히 실장 하는 장비이다. 이를 위해서는 중간 위치에 설치된 카메라 앞에서 정지하여 비전 검사를 한 후 실장 위치로 이동한다. 그 이동 경로의 타입은 16가지가 되며, 그 중에서 (sLa-Camera-pLb)경로에 대해 가장 빠른 시간에 실장 할 수 있는 알고리즘을 본 논문에서는 제시한다. 이를 위해서는 카메라 위치에서 정지하지 않고 이동하면서 비전 검사할 때 수많은 궤적 중 어떤 궤적이 가장 빠른 시간에 검사 후 실장 할 수 있는 경로인지를 찾는 알고리즘을 제시한다. 그 결과 카메라 앞에서 정지하여 검사 후 이동하는 방식보다 16%의 시간 단축 효과가 있음을 시뮬레이션을 통해 입증하였다.

SMT is an equipment that picks up electronic components and does precise placing onto PCBs. In order to do this, it stops in front of a camera installed in the middle to go over vision inspection. And after that it is move for placing. There are 16 different types of routes in this process. This paper presents the fastest algorithm to place (sLa-Camera-pLb) among all these routes. In order to do this, instead of stopping in front of camera the object should move on while going over the vision inspection. Among all possible tracks, this thesis will provide algorithm to find out the fastest tracks to do vision inspection and placing. And as a result, this thesis have demonstrated that this method can save about 16% of time compared to going over inspection while the object is standing still through simulation.

글로벌 기업들은 글로벌 생산 네트워크를 구축하여 세계 시장의 수요에 대응하고 있다. 그러나 해외 공장 설립 및 운영 시 현지 노동자들의 임금문제, 낮은 생산의욕 및 도난사고로 인한 피해가 속출하고 있다. 따라서 본 논문에서는 글로벌 기업의 해외 공장에서 원자재를 IP Camera와 RFID Box를 이용하여 관리함으로써, 실시간적인 재고 파악과 원 자재의 도난 또는 멸실되는 상황을 방지할 수 있는 원자재 유실 방지 모델을 제안하였다. 이 모델은 원격지에 있는 재 고의 관리가 가능하여 보다 비용 효율적인 생산관리를 실현할 수 있게 해주며, 불합격품 거래로 인한 브랜드 이미지 손 실과 수익 저하를 방지할 수 있도록 해준다.

Global companies are building global production network for the demand of the global market. However, the establishment and operation of overseas plants make the unexpected case like theft, low will to produce and salary issue of local workers. Thus in this paper, we propose the raw materials loss prevention model based on IP Camera and RFID Box for raw materials management of overseas plants. This model can prevent the theft or loss of raw materials write inventory up in real-time. This model can be allows us to realize the cost-effective production management because it enables remote inventory management. And this model can prevent brand image danage and profit loss due to reject product.

Even if some kinds of software for halftoning such as raster image process program are on the market, most of them couldn`t apply to the new laser gravure cylinder making machine because they are closed the source program. In order to develop new laser gravure engraving machine, we were investigated new suitable halftoning software based on the mapping and equal density expansion algorithm that we are mapped. As the results of actual practice with Microsoft Visiul C++2.0 in 133Khz 64M ram IBM personal computer, we found that the new software for halftone generation for laser gravure engraving machine are available in new machine, and easily load and processing the graphic image for gravure hardcopy.

본 연구는 6 Dof(Degrees of freedom) 지원 가상현실 재현을 위한 3차원 기하정보 기반 Mesh Data 변 환 및 CG/실사 기반 영상 합성 기술을 통해 공간콘텐츠를 효과적으로 제작할 수 있는 새로운 방안을 제안한다. 기존 가상현실 콘텐츠는 CG기반 콘텐츠 제작 기술이 주로 활용되었으나 사실적인 실감콘텐츠 구현을 위해서는 실사 기반 3차원 공간 구성이 필수적이다. 본 논문에서는 2D Arrary 영상을 기반으로 전경 및 배경 영상을 3차 원 공간콘텐츠로 구성하여 실사 기반 실감형 가상현실 콘텐츠를 제작하였다. 실측 공간과 가상현실 공간 정합도를 계산하여 3차원 공간 구성의 정확도를 측정하였다. 향후 LF(Light Field) 영상 및 디지털 트윈 기술을 적용하여 실감형 가상현실 콘텐츠를 제작하는데 활용 수 있다.

This study proposes a new method to effectively create spatial contents through 3D geometry based mesh data transformation and CG / real image based synthesis technology for 6 Dof (Degrees of freedom) support. Conventional virtual reality contents are mainly used for CG-based content production technology, but due to the real-based 3D space construction is essential for realizing realistic contents. Based on the 2D Arrary image, the foreground and background images were composed of 3D spatial contents to produce photo realistic virtual reality contents. The accuracy of three-dimensional spatial composition was measured by calculating the degree of matching between the measured space and the virtual reality space. In the future, it can be used to produce realistic virtual reality contents by applying LF image and digital twin technology.

In this paper we present a system that enables the user to interact with large display system even without touching the screen. With two infrared sensitive cameras mounted on the bottom left and bottom right of the display system pointing upwards, the user fingertip position on the selected region of interest of each camera view is found using vertical intensity profile of the background subtracted image. The position of the finger in two images of left and right camera is mapped to the display screen coordinate by using pre-determined matrices, which are calculated by interpolating samples of user finger position on the images taken by pointing finger over some known coordinate position of the display system. The screen is then manipulated according to the calculated position and depth of the fingertip with respect to the display system. Experimental results demonstrate an efficient, robust and stable human computer interaction.

Camera calibration is an important part of machine vision and close-range photogrammetry. Since current calibration methods fail to obtain ideal internal and external camera parameters with limited computing resources on mobile terminals efficiently, this paper proposes an improved fast camera calibration method for mobile terminals. Based on traditional camera calibration method, the new method introduces two-order radial distortion and tangential distortion models to establish the camera model with nonlinear distortion items. Meanwhile, the nonlinear least square L-M algorithm is used to optimize parameters iteration, the new method can quickly obtain high-precise internal and external camera parameters. The experimental results show that the new method improves the efficiency and precision of camera calibration. Terminals simulation experiment on PC indicates that the time consuming of parameter iteration reduced from 0.220 seconds to 0.063 seconds (0.234 seconds on mobile terminals) and the average reprojection error reduced from 0.25 pixel to 0.15 pixel. Therefore, the new method is an ideal mobile terminals camera calibration method which can expand the application range of 3D reconstruction and close-range photogrammetry technology on mobile terminals.

This paper presents a scalable multiple camera collaboration strategy for active tracking applications in large areas. The proposed approach is based on distributed mechanism but emulates the master-slave mechanism. The master and slave cameras are not designated but adaptively determined depending on the object dynamic and density distribution. Moreover, the number of cameras emulating the master is not fixed. The collaboration among the cameras utilizes global and local sectors in which the visual correspondences among different cameras are determined. The proposed method combines the local information to construct the global information for emulating the master-slave operations. Based on the global information, the load balancing of active tracking operations is performed to maximize active tracking coverage of the highly dynamic objects. The dynamics of all objects visible in the local camera views are estimated for effective coverage scheduling of the cameras. The active tracking synchronization timing information is chosen to maximize the overall monitoring time for general surveillance operations while minimizing the active tracking miss. The real-time simulation result demonstrates the effectiveness of the proposed method.

Objective: This study aimed to evaluate the validity of a depth camera-based system for monitoring physical function, assessing its feasibility for accurately monitoring activities of daily living. Design: A cross-sectional study. Methods: Twenty-three participants were enlisted to perform fifteen activities of daily living within a living laboratory designed to simulate a home environment. Activities were monitored using a depth camera system capable of classifying actions into standing, sitting, and lying down, with a conventional video camera employed for activity recording. The duration of each activity, as measured by the system, was compared to direct observations made by a physical therapist which were analyzed using a motion analysis software. The association between these two measurement approaches was assessed through correlation analysis, coefficient of determination, intraclass correlation coefficient (ICC), and Bland-Altman plots. Results: Our findings indicated that standing activities exhibited the highest correlation (r=0.847) between the system measurements and physical therapist observations, followed by sitting (r=0.817) and lying down (r=0.734), which demonstrated lower correlations. However, the ICC and Bland-Altman plots revealed notable variances between the two measurement methods, particularly for activities involving lying down. Conclusions: In this study, the depth camera-based physical monitoring system showed promise feasibility in distinguishing standing, sitting, and lying down activities at home environments. However, the current study also underlined some necessities of enhancements in capturing lying down activities.

Throughout the last decade, a large number of electronic driver-assistance systems have been developed in order to improve vehicles' comfort and safety. Up to now, they more or less all rely on odometric data. However in the future, a new form of information is likely to be used: visual sensors, e.g. cameras, can provide additional knowledge on traffic and the vicinity. This paper presents an inertially stabilized camera platform for high precision tracking of points of interest, that has been developed within the project FORBIAS (Research cooperation for biomimetic assistance systems). The inertial stabilization is based on a specially designed sensor system that uses MEMS components in connection with a biologically inspired control algorithm for adaptive enhancement and online calibration of the gaze stabilization system.

최근 인공지능 기술과 인공지능을 처리할 수 있는 엣지AI단말의 발전으로, 다양한 영상 분석 서비스에 소프트웨어 정의 카메라(Software-defined Camera)가 활발히 활용되고 있다. 그러나 적용되는 소프트웨어 정의 카메라의 수 가 늘어날수록 소프트웨어 정의 카메라 자체 뿐만 아니라 핵심 구성요소인 엣지AI단말의 관리의 어려움과 비용 문제 가 대두되고 있다. 따라서 본 논문에서는 엣지AI단말의 운영 효율성과 엣지AI단말에서 제공하는 서비스의 성능을 향 상시키기 위한 OTA(Over-the-AIr) 기반의 엣지AI단말 관리 시스템을 제안하고 구현한다. 제안한 시스템은 AI학 습모델과 서비스 애플리케이션의 원격 업데이트 및 관리 기능을 제공함으로써 물리적인 접근 없이도 엣지AI단말을 효율적으로 관리할 수 있도록한다. 또한, 제안한 시스템은 엣지AI단말의 상태정보를 주기적으로 수집하여 실시간으 로 운영상태를 모니터링할 수 있으며, 앞서 언급한 기능들을 Web 기반의 사용자 친화적인 GUI를 통해 제공함으로 써 통합적으로 관리할 수 있도록 한다. 제안한 시스템은 Nvidia Jetson Orin Nano 기반으로 엣지AI단말과 데스크 탑 기반으로 통합관제서버를 구현하였으며, 실증을 통해 설계한 기능들이 정상적으로 동작함을 확인하였다.

With the recent advancements in artificial intelligence(AI) technology and the development of edge AI devices capable of processing AI workloads, software-defined cameras(SDCs) have been widely utilized in various video analysis services. However, as the number of deployed SDCs continues to grow, the complexity and cost of managing not only the cameras themselves but also their core components-edge AI devices-have emerged as significant challenges. To address these issues, this paper proposes and implements an Over-the-AIr(OTA)-based edge AI device management system designed to improve the operational efficiency of edge AI devices and enhance the performance of the services they provide. The proposed system supports remote updates and management of AI models and service applications, enabling efficient mAIntenance without requiring physical access to the devices. Furthermore, it periodically collects device status information to support real-time monitoring of operational conditions and provides all management features through an integrated, user-friendly Web-based graphical user interface(GUI). The system was implemented using Nvidia Jetson Orin Nano-based edge AI devices and a desktop-based integrated control server. Functional verification through practical deployment confirmed the successful operation of all designed components.