• The Journal of Korea Robotics Society
• /
• v.14 no.1
• /
• pp.1-7
• /
• 2019

Recently, smart factories have attracted much attention as a result of the 4th Industrial Revolution. Existing factory automation technologies are generally designed for simple repetition without using vision sensors. Even small object assemblies are still dependent on manual work. To satisfy the needs for replacing the existing system with new technology such as bin picking and visual servoing, precision and real-time application should be core. Therefore in our work we focused on the core elements by using deep learning algorithm to detect and classify the target object for real-time and analyzing the object features. We chose YOLO CNN which is capable of real-time working and combining the two tasks as mentioned above though there are lots of good deep learning algorithms such as Mask R-CNN and Fast R-CNN. Then through the line and inside features extracted from target object, we can obtain final outline and estimate object posture.

• The KIPS Transactions:PartB
• /
• v.13B no.7 s.110
• /
• pp.663-670
• /
• 2006

For virtual target to be displaying on real image realistically, it is essential to determine the location of the virtual object together with producing the occlusions area after registering two world. In this paper, we propose the new method to solve occlusions which happens during virtual target moves according to the simulated route on real image. For this purpose, we first construct three dimensional virtual world by DEM of experimental area and register CCD camera image on it by visual clues. Next, we also propose a method to solve the occlusion using snake and picking algorithm which can extract the three dimensional information of the position happening occlusion in the image and can use it when target moves that area. In the experiment, we proved the effectiveness of the proposed method in the environment which a partial occlusions happens.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.18 no.2
• /
• pp.206-211
• /
• 2008

Trainee can enhance his perception of and interaction with the real world by displayed virtual objects in simulation using image processing technology. Therefore, it is essential for realistic simulation to determine the occlusion areas of the virtual object produces after registering real image and virtual object exactly. In this paper, we proposed the new method to solve occlusions which happens during virtual target moves according to the simulated route on real image using improved object contour extraction by neighboring edge search and picking algorithm. After we acquire the detailed contour of complex objects by proposed contour extraction algorithm, we extract the three dimensional information of the position happening occlusion by using MER for performance improvement. In the experiment, we compared proposed method with existed method and preyed the effectiveness in the environment which a partial occlusions happens.

• The Journal of Korea Robotics Society
• /
• v.3 no.1
• /
• pp.73-80
• /
• 2008

Many researchers are studying on humanoid robots in all over the world. However the humanoid robots are still limited in doing works like picking up objects on the ground or moving rapidly. In this study, a humanoid robot based on the wheel-driving was developed. It can operate with a human working area keeping the stability. Also, the developed robot can take up the object on the floor since it has knee(1DoF) and waist(3DoF), and do service quickly and steadily. The hardware and software structure and algorithms of the developed robot, SEROPI are introduced in this paper.

• Annual Conference of KIPS
• /
• 2021.11a
• /
• pp.600-601
• /
• 2021

식품제조 현장에서 유연포장을 수행하기 위해, 로봇이 다양한 제품들을 파지하거나 이송하는 작업이 필수적이다. 본 논문에서는 전통장류 식품제조에서 다양한 종류와 무게를 조합하는 혼합포장을 위해 전통장류 물체 영역 인식과 종류 인식을 수행한다. 이를 위하여, 대표적인 전통장류에 대해 종류를 분류하고, RGB-D 데이터를 입력으로 물체 영역과 종류 인식을 수행하는 딥러닝 네트워크를 학습한다. 실험 결과를 통해, 물체 영역의 중심점을 기반으로 흡착 기반 파지점을 선정할 수 있음을 확인한다.

• Journal of Broadcast Engineering
• /
• v.26 no.6
• /
• pp.704-713
• /
• 2021

This paper proposes a method to prioritize the virtual objects to be selected, considering both the user's hand and the geometric relationship with the virtual objects and the user's intention which is recognized in advance. Picking up virtual objects in VR content is an essential and most commonly used interaction. When virtual objects are located close to each other in VR, a situation occurs in which virtual objects that are different from the user's intention are selected. To address this issue, this paper provides different weights for user intentions and distance between user's hand and virtual objects to derive priorities in order to generate interactions appropriately according to the situation. We conducted the experiment in the situation where the number of virtual objects and the distance between virtual objects are diversified. Experiments demonstrate the effectiveness of the proposed method when the density between virtual objects is high and the distance between each other is close, user satisfaction increases to 20.34% by increasing the weight ratio of the situation awareness. We expect the proposed method to contribute to improving interaction skills that can reflect users' intentions.

• International Journal of Aeronautical and Space Sciences
• /
• v.18 no.4
• /
• pp.740-756
• /
• 2017

This paper proposes a feedback-linearization-based control algorithm for multirotor unmanned aerial vehicles (UAVs). The feedback linearization scheme is highly efficient for considering nonlinearity between the rotational and translational motion of multirotor UAVs. We also propose a dynamic equation that reflects the aerodynamic effects of the vehicles; the equation's parameters can be determined through curve fitting using actual flight data. We derive the feedback linearization controller from the proposed dynamic equation, and propose a Luenberger observer to attenuate measurement noises. The proposed algorithm is implemented using our in-house flight control computer, and we describe its implementation in detail. To investigate the performance of the proposed algorithm, we carry out two flight scenarios: the first scenario, an autonomous landing on a moving platform, is a test of maneuverability; the second, picking up and replacing an object, test the algorithm's accuracy. In these scenarios, the proposed algorithm precisely controls multirotor UAVs, and we confirm that it can be successfully applied to real flight environments.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.6 no.6
• /
• pp.797-804
• /
• 2002

As internet use is being generalized central of WWW(World Wide Web) service of multimedia based recently, we could acquire many informations that exist to all over the world's computer network .Therefore, picking up of information became important problem before that internet is generalized, but it is risen to important problem to acquire correct information rapidly on modem society that use of internet is generalized. This paper designed internet search engine and understand structure of internet search engine drawing URL that is optimized, and secure embodiment technology using Java that is language of object base. Search engine that proposed in this paper maintained user's the convenience by offer keyword search, and simplify user interface And although quantity of searched information site is few, search engine show that the bad link rate of searched result is improved compare with existent domestic manufacture search engines.

• 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.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.47 no.2
• /
• pp.10-20
• /
• 2024

In recent automated manufacturing systems, compressed air-based pneumatic cylinders have been widely used for basic perpetration including picking up and moving a target object. They are relatively categorized as small machines, but many linear or rotary cylinders play an important role in discrete manufacturing systems. Therefore, sudden operation stop or interruption due to a fault occurrence in pneumatic cylinders leads to a decrease in repair costs and production and even threatens the safety of workers. In this regard, this study proposed a fault detection technique by developing a time-variant deep learning model from multivariate sensor data analysis for estimating a current health state as four levels. In addition, it aims to establish a real-time fault detection system that allows workers to immediately identify and manage the cylinder's status in either an actual shop floor or a remote management situation. To validate and verify the performance of the proposed system, we collected multivariate sensor signals from a rotary cylinder and it was successful in detecting the health state of the pneumatic cylinder with four severity levels. Furthermore, the optimal sensor location and signal type were analyzed through statistical inferences.