Earticle

스마트폰의 보급과 이를 활용한 SNS의 사용자가 증가하면서, 기존의 데이터 타입인 텍스트와 같은 정형 데이터뿐만 아니라 이미지, 동영상과 같은 비정형 데이터가 실시간으로 엄청난 양으로 발생하고 있다. 이런 비정형 데이터를 활용한 서비스가 늘어나면서 다양한 솔루션들이 출시되고 있다. 기존의 정형 데이터는 텍스트 분석이나 간단한 질의를 통해 분석 및 원하는 결과를 쉽게 얻을 수 있지만, 동 영상과 같은 비정형 데이터는 더욱 복잡한 과정을 통해서 분석되어야만 한다. 본 연구에서는 라즈베리파이에 연결된 인체감지 센서와 카메라 센서를 통해 수집한 영상 데이터를 전송받아 특정한 객체를 검출하는 방안을 제시하고자 한다. 이를 위해 관련 연구로서 라즈베 리파이의 하드웨어적인 특성과 컨볼루션 신경망 이론에 대해서 조사하였고, 이를 바탕으로 라즈베리파이 보드에 인체감지 센서와 카 메라 센서로 구성된 카메라 모듈을 구성하여 주변을 촬영한 영상을 관리 서버에 전송한 다음에 이를 YOLO 알고리즘을 사용해서 특정 한 객체를 검출하는 방안을 기술한다. 이를 통해 보다 쉬운 방식으로 객체를 검출함으로써 보다 다양한 분야에 응용할 수 있다.

With the spread of smartphones and the increasing number of users of SNS using them, not only structured data such as text, which is an existing data type, but also unstructured data such as images and videos are generated in huge amounts in real time. As services using such unstructured data increase, various solutions are being released. Existing structured data can be easily analyzed and desired results obtained through text analysis or simple queries, but unstructured data such as video must be analyzed through a more complicated process. In this study, we propose a method of detecting a specific object by receiving image data collected through a human body sensor and camera sensor connected to Raspberry Pi. To this end, as a related study, the hardware characteristics of Raspberry Pi and the theory of convolutional neural networks were investigated. On that basis, a method of detecting a specific object using the YOLO algorithm is described. Through this, it can be applied to more diverse fields by detecting objects in an easier way.

Vision, Radar, and LiDAR sensors are widely used for autonomous vehicle perception technology. Especially object detection and classification are primarily dependent on vision sensors. However, under poor lighting conditions, dazzling sunlight, or bad weather an object might be difficult to be identified with general vision sensors. In this paper, we propose a sensor fusion system that combines a thermal infrared camera and a LiDAR sensor that can reliably detect and identify objects even in environments with poor visibility, such as day or night. The proposed method obtains the external parameters of the two sensors by designing and manufacturing a 3D calibration target to externally calibrate the thermal infrared camera and the LiDAR sensor. To verify the performance, experiments were conducted in day and night environments. The proposed sensor system and fusion algorithm show that it can reliably detect and identify objects even in environments with poor visibility, such as day or night.

This study proposes an environment-adaptive sensor weight fusion model for autonomous driving systems that automatically optimizes the weights between LiDAR and camera sensors based on environmental conditions. Using an F1TENTH autonomous driving platform, we implemented a dual-branch neural network that processes data from 2D LiDAR and monocular camera sensors and combines their outputs with learnable weights. Through extensive experiments in various indoor environments, we found that the optimal sensor weight ratio in general conditions was approximately 64% for LiDAR and 36% for camera sensors. However, these weights dynamically adjusted based on environmental changes, with LiDAR reliance increasing in low-light conditions and camera influence growing in well-lit environments. Our results highlight the importance of dynamic sensor fusion in autonomous driving systems and provide insights into optimal weight distributions across different driving scenarios. Future work will extend this approach to larger-scale platforms with additional sensors such as GPS.

유비쿼터스 환경에서 최근 스마트 폰에서 증강현실(Augmented Reality) 서비스가 활발히 구현되고 있다. 본 논문에서는 아이폰에 탑재된 카메라와 가속도 센서를 이용한 모바일 증강현실 시스템을 개발하였다. 일반적으로 카메라만을 이용하는 경우, 가려짐이나 빠른 움직임으로 인해 마커를 검출하지 못하는 상황이 발생한다. 이를 해결하기 위해 마커가 가려지는 시점에서부터 아이폰에서 제공되는 가속도 출력값을 사용하여 일정 시간 동안의 카메라의 움직임을 해석함으로써 모바일 게임용 증강현실 어플리케이션을 제공하는 기술이 제안된다. 실험결과로부터 마커가 가려지는 경우에도 모바일 단말의 움직임을 효과적으로 해석함을 확인하였다.

Augmented Reality (AR) services on a smart phone have been developed actively in ubiquitous environment. This paper presents a novel mobile system using camera and accelerometer sensor built in iPhone. Using the camera sensing disables to deal with the case the marker for AR is occluded or the camera moves so fast. By using the output of the acceleration sensing in iPhone, the proposed method provides continuous AR application even though when the marker is undetected. The experimental results showed that we achieve a precise estimation of the device movement in AR application such as mobile game.

본 논문에서는 적외선 열화상 센서와 RGB 센서를 통합한 멀티모달 지능형 카메라의 설계 및 개발을 제안하였다. 제안 시스템은 두 센서를 단일 프레임 구조로 통합하여 온도 정보와 영상 데이터를 시간 동기화 방식으로 수집하며, TensorFlow Lite 기반의 경량 Edge AI 모델을 적용하여 실시간 위험 감지를 수행하였다. 8-bit 양자화와 경량 네트워크 구조를 적용하여 저전력 환경에서도 안정적인 추론이 가능하도록 설계하였다. 또한, 열화상과 RGB 데이터의 융합 처리를 통해 조명 변화나 연기 등 환경적 제약 조건에서도 감지 신뢰성을 향상시켰다. 실험결과, 제안된 열화상–RGB 센서 융합 방식은 단일 센서 기반 시스템 대비 향상된 위험 감지 성능과 환경 적응성을 보였다.

This paper presents the design and development of a multimodal intelligent camera integrating thermal and RGB sensors. The proposed system synchronizes temperature and visual data within a unified hardware frame and performs real-time hazard detection using a lightweight TensorFlow Lite-based edge AI model. To enable low-power operation, 8-bit quantization and lightweight neural network architectures are applied. Furthermore, the fusion of thermal and RGB modalities enhances detection reliability under challenging environmental conditions such as low illumination or smoke. Experimental results demonstrate that the proposed thermal–RGB sensor fusion approach improves detection performance and environmental robustness compared to single-modality systems.

최근 UHDTV와 같은 고품질, 고화질을 기본으로 현장감을 제공하는 방송 디스플레이 환경이 요구됨에 따라 다양한 4K 카메라들이 출시되고 있다. 일반적으로 사용되고 있는 Super 35mm 센서의 4K/UHD 카메라는 광고, 드라마, 자연다큐멘터리, 교양 등 싱글 카메라 제작 시스템에는 활용도가 높다. 하지만 예능, 스포츠, 콘서트 등 멀티카메라 제작시스템에서는 거리에 따라 클로즈 업(close-up)이나 타이트 샷(tight shot)의 한계점을 가짐에 따라 방송 중계에 제약이 따른다. 또한 4K 렌즈는 기존의 HD 렌즈 군에 비해 부족한 실정이라 HD 제작 시 구현되었던 카메라 촬영 샷에 제한이 따른다. 따라서 이를 보완하기 위한 2/3inch 센서 4K/UHD 카메라의 활용이 요구되고 있으며, 다양한 렌즈군의 활용으로 UHD 영상 콘텐츠 제작이 요구되고 있다. 본 연구는 Grass Valley 사의 LDX 4K/UHD카메라를 사용하여 2/3" 센서에서의 HD급 렌즈 활용에서의 화질 테스트를 통해 활용 가능성에 대해 분석하고, 이를 효율적인 활용 가능성을 제시하였다.

Recently, there is more and more demand on broadcasting display environment that provides a sense of realism basically with high quality and high definition as in UHDTV, so various kinds of 4K cameras have been released. Super 35mm sensor 4K/UHD cameras being generally used can be usefully applied to the single camera-based system such as advertisements, dramas, nature documentaries, and culture programs. However, in the multicamera-based system such as entertaining programs, sports, and concerts, relay broadcasting may face limitations as there are restrictions of close-up or tight shots according to the distance. Besides, 4K lenses are not enough compared with the previous HD lenses' group, so there are limitations in camera photographing shots that have been realized in HD filming. Therefore, to complement this, the use of 2/3 inch sensor 4K/UHD cameras is being demanded, and it is also needed to use various lenses' groups to produce UHD image contents. Accordingly, this study used Grass Valley's LDX 4K/UHD camera to analyze its applicability based on the picture quality test at the use of HD lenses in 2/3 sensors and also the possibility of its efficient application.

Due to the variety of signal processing and complicated mathematical analysis, it is not easy to accomplish 3D bin-picking with non-contact sensor. To solve this difficulties the reliable signal processing algorithm and a good sensing device has been recommended. In this research, 3D laser scanner and CCD camera is applied as a sensing device respectively. With these sensor we develop a two-step bin-picking method and reliable algorithm for the recognition of 3D bin object. In the proposed bin-picking, the problem is reduced to 2D intial recognition with CCD camera at first, and then 3D pose detection with a laser scanner. To get a good movement in the robot base frame, the hand eye calibration between robot's end effector and sensing device should be also carried out. In this paper, we examine auto-calibration technique in the sensor calibration step. A new thinning algorithm and constrained hough transform is also studied for the robustness in the real environment usage. From the experimental results, we could see the robust bin-picking operation under the non-aligned 3D hole object.

본 논문은 환경변화에 강인한 센서 융합 교통정보 수집 시스템을 제안한다. 제안된 시스템은 각 센서의 특징을 융합하여 영상 검지기에 비해 환경에 강인하고, 주·야간 등 시간에영향을 받지 않으며 루프 검지기에 비해 유지보수에 드는 비용이 적다. 이는 레이더의 물체추적 기법과 영상검지기의 차량분류, 적외선거리계의 신뢰성 높은 객체검지 정보를 융합하고 각 센서의 문제점을 보완하여 개선을 이루었다. 구현된 시스템을 보행자의 통행이 가능한 도로에서 주·야간 5일에 걸쳐 6시간 동안 실험한 결과 88.7%의 분류정확도와 95.5%의 차량 검지율을 나타내었다. 본 시스템의 파라미터 최적화 작업을 실험환경에 따라 적응되는 방식으로 보완한다면 교통정보 수집 체계의 고도화에 기여할 수 있을 것으로 기대된다.

The purpose of this paper is to develop a multi-sensor fusion-based traffic information acquisition system with robust to environmental changes. it combines the characteristics of each sensor and is more robust to the environmental changes than the video detector. Moreover, it is not affected by the time of day and night, and has less maintenance cost than the inductive-loop traffic detector. This is accomplished by synthesizing object tracking informations based on a radar, vehicle classification informations based on a video detector and reliable object detections of a infrared range finder. To prove the effectiveness of the proposed system, I conducted experiments for 6 hours over 5 days of the daytime and early evening on the pedestrian - accessible road. According to the experimental results, it has 88.7% classification accuracy and 95.5% vehicle detection rate. If the parameters of this system is optimized to adapt to the experimental environment changes, it is expected that it will contribute to the advancement of ITS

제조 패러다임 변화에 따라 다양한 협동로봇이 신규시장을 창출하고 있다. 협동로봇은 기존 산업용 로봇 대비 쉬운 운용, 생산성 향상, 단순 작업을 하는 인력을 대체하는 목적으로 모든 산업에서 협동로봇의 수요가 증가하고 있다. 그러나 산업현장에서 협동로봇으로 인한 작업 중 사고가 빈번하게 발생하고 있으며, 작업자의 안전을 위협하고 있다. 인간 중심의 환경에서 로봇을 통한 산업현장이 구성되려면 작업자의 안전을 보장해야 하며 출동 가능성을 업애고 신뢰할 수 있는 통신을 하는 협동로봇 가드 시스템의 개발의 필요성이 있다. 센서 및 컴퓨터 비전을 통해 협동로봇의 작업 반경 내에서 발생하는 사고를 이중으로 방지하고 안전사고 위험을 감소시켜야 한다. 다양한 산업용 장비와 통신을 위한 국제 프로토콜인 OPC UA를 기반으로 시스템을 구축하고 초음파 센서와 CNN(Convolution Neural Network)적용한 영상 분석을 통한 협동로봇 가드 시스템을 제안한다. 제안 된 시스템은 작업자의 불안전한 상황에서 로봇 제어의 가능성을 평가한다.

The robot is the creation of new markets and various cooperation according to the manufacturing paradigm shift. Cooperative management easy for existing industrial robots, robots work on productivity, manpower to replace the robot in every industry cooperation for the purpose of and demand increases.to exist But the industrial robot at the scene of the cooperation working due to accidents are frequent, threatening the safety of the operator. Of industrial site is configured with a robot in an environment ensuring the safety of the operator to and confidence to communicate that can do the possibility of action.Robot guard system of the need for development cooperation. The robot's cooperation through the sensors and computer vision task within a radius of the double to prevent accidents and accidents should reduce the risk. International protocol for a variety of industrial production equipment and communications opc ua system based on ultrasonic sensors and cnn to (Convolution Neural Network) for video analytics.We suggest the cooperation with the robot guard system. Robots in a proposed system is unsafe situation of workers evaluating the possibility of control.

Occupancy detection systems are crucial for optimizing energy efficiency in smart cities and buildings but often face privacy and data dependency challenges. YOLO (you only look once), a widely used real-time detection framework, relies on identifiable image data and labeled datasets. This study proposes a privacy-preserving, labeling-free occupancy sensor using a time-of-flight (ToF) camera, and a clustering algorithm. Positioned above doorways, the ToF camera captures depth data that inherently protect privacy by avoiding identifiable information. Using the mean shift clustering algorithm, it performs real-time detection and tracking without labeled data, generating bounding boxes for movement analysis. Unlike traditional ToF-based or unsupervised methods, the proposed system adapts dynamically to varying occupant behaviors and environmental conditions for robust real-time detection. Experimental results show that the proposed method achieves over 90% accuracy in standard single-entry and exit scenarios. By addressing existing limitations, it offers a data-efficient, privacy-sensitive solution for building digital twins in energy optimization and resource management.

This paper proposes a sensor-fusion technique where the data sets for the previous moments are properly transformed and fused into the current data sets to enable accurate measurement, such as, distance to an obstacle and location of the service robot itself. In the conventional fusion schemes, the measurement is dependent on the current data sets. As the results, more of sensors are required to measure a certain physical parameter or to improve the accuracy of the measurement. However, in this approach, instead of adding more sensors to the system the temporal sequence of the data sets are stored and utilized for the measurement improvement. Theoretical basis is illustrated by examples and the effectiveness is proved through the simulations. finally, the new space and time sensor fusion (STSF) scheme is applied to the control of a mobile robot in an unstructured environment as well as structured environment.

Camera rigs for shooting 3D video are classified as manual, motorized, or fully automatic. Even in an automatic camera rig, the process of Stereoscopic 3D (S3D) video capture is very complex and time-consuming. One of the key time-consuming operations is capturing the distance parameters, which are near distance, far distance, and convergence distance. Traditionally these distances are measured by tape measure or triangular indirect measurement methods. These two methods consume a long time for every scene in shot. In our study, a compact laser distance sensing system with long range distance sensitivity is developed. The system is small enough to be installed on top of a camera and the measuring accuracy is within 2% even at a range of 50 m. The shooting time of an automatic camera rig equipped with the laser distance sensing system can be reduced significantly to less than a minute.