• 대한전자공학회논문지
• /
• 제27권12호
• /
• pp.1817-1824
• /
• 1990

This paper describes a shape matching algorithm for occluded or distorted two-dimensional objects. In our approach, the shape matchin is viewed as a segment matching problem. A shape matching algorithm, based on both the stochastic labeling technique and the hypothesis generate-test paradigm, is proposed, and a simple technique which performs the stochastic labeling process in accordance with the definition of consisten labeling assignment without requiring an iterative updating process of probability valiues is also proposed. Several simulation results show that the proposed algorithm is very effective when occlusion, scaling or change of orientation has occurred in the object.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1991년도 한국자동제어학술회의논문집(국내학술편); KOEX, Seoul; 22-24 Oct. 1991
• /
• pp.990-995
• /
• 1991

The integration of intelligent robots into manufacturing systems should positively impact the product quality and productivity. A new theory of object location and recognition using the supporting plane is presented. The unknown supporting points are determined by image coordinates, known camera parameters, and joint coordinates of the robot manipulators. This is developed by using the geometrical interpretation of perspective projection and the geometrical constraints of industrial environments. This can be applied to solve typical robot vision problems such as determination of position, orientation, and recognition of objects.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1992년도 한국자동제어학술회의논문집(국내학술편); KOEX, Seoul; 19-21 Oct. 1992
• /
• pp.1014-1017
• /
• 1992

This paper proposed LM method for solving bin-picking problem in robot vision. It has the processing steps such as image enchancement, image thresholding, region labelling, and moment computation. To determine a target object form bined objects, the modified labelling method is used. To determine position and orientation of holdsite, the moment method is used. Finally, some experiment results are illustated and compared with the results of conventional shrinking algorithm. The proposed LM method has reduced processing time.

• 대한기계학회논문집A
• /
• 제40권9호
• /
• pp.783-790
• /
• 2016

IMU기반 자세결정에 있어 추정 정확도의 저하요인 중 주요한 한 가지는 운동체의 가속도이다. 이는 가속도가 크게 발생하는 경우 가속도계 신호는 더이상 수직축 참조벡터가 될 수 없기 때문이다. 이에 대한 대책으로 일부 자세추정 알고리즘에서는 가속도 보상 메카니즘이 적용되어 왔다. 가장 보편적이고 간단한 스위칭 방법부터 적응추정방식, 가속도 모델기반 방식 등이 제안되어 왔으나, 이들 보상 메카니즘에 대한 비교분석은 이루어 지지 않았다. 본 논문은 쿼터니언기반의 Pseudo 칼만필터를 바탕으로 하여 세 가지 가속도 보상 메커니즘의 성능을 비교분석하였다. 가속조건 실험 분석을 통해 다음을 확인할 수 있었다. (1) 가속구간에서의 추정정확도 저하를 방지하기 위해선 가속도 보상 메카니즘이 반드시 필요하다. (2) 단순 스위칭 방법도 상당한 효과를 보였으나, 보다 정교한 적응추정 방식과 가속도 모델방식이 동등수준으로 가장 정확한 결과를 보였다.

• 한국CDE학회논문집
• /
• 제16권1호
• /
• pp.1-10
• /
• 2011

In this paper, we propose an AR-based tangible interaction using a finger fixture for virtual evaluation of digital handheld products. To realize tangible interaction between a user and a product in a computer-vision based AR environment, we uses two types of tangible objects: a product-type object and a finger fixture. The product-type object is used to acquire the position and orientation of the product, and the finger fixture is used to recognize the position of a finger tip. The two objects are fabricated by RP technology and AR markers are attached to them. The finger fixture is designed to satisfy various requirements with an ultimate goal that the user holding the finger fixture in his or her index finger can create HMI events by touching specified regions (buttons or sliders) of the product-type object with the finger tip. By assessing the accuracy of the proposed interaction, we have found that it can be applied to a wide variety of digital handheld products whose button size is not less than 6 mm. After performing the design evaluation of several handheld products using the proposed AR-based tangible interaction, we received highly encouraging feedback from users since the proposed interaction is intuitive and tangible enough to provide a feeling like manipulating products with human hands.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 1997년도 추계학술대회 논문집
• /
• pp.1067-1071
• /
• 1997

A new type of gripper for micrometer-size objects is developed using piezoelectric multi-layer benders. It is composed of three chopsticks, two of which are designed to grip micro-objects. The third one is reserved for helping the two when objects are released from the chopsticks. It is well known that a micro object is much easier to grasp than to release it after holding it. The electrostatic force between the chopsticks and an object is believed to be the main cause of adhesion in a dry environment. The surface tension becomes very important when liquids are present or in a liquid. The third auxiliary chopsticks is introduced to solve there surface effects. All the three chopsticks are made of tungsten wires with sharpened ends by etching. When grasping microparts, the two chopsticks are utilized, and, when releasing them anywhere the parts are located, the third one reduces the electrostatic force between the objects and the chopstick may be to help the other two chopsticks to hold an objects in a desired orientation. We constructed the three chopstick gripoer for micro objects and test their function by holding and releasing an object of a diameter of 100 micrometers. We make use of open loop voltage control. The bender displacement resolution is sub-micrometer. The gripping forces, about tens of mN are obtained. The experiment shows that the third auxiliary chopstick functions effectively.

• 대한전자공학회논문지SP
• /
• 제38권1호
• /
• pp.89-96
• /
• 2001

본 논문에서는 잡음에 강건한 일반화 대칭 변환을 주목 연산자로 제안하고 이를 이용하여 크기와 형태가 다양한 물체들을 효과적으로 검출하였다. 기존의 주목 연산자와는 달리 두 화소의 명도변화의 크기와 대칭성뿐만 아니라 방사(radial)방향 명도변화의 수렴 및 발산을 누적 대칭도에 반영시킴으로써 명도변화 방향의 일관된 수렴이나 발산이 없는 잡음 영역에 의한 대칭 기여도가 누적되지 않도록 하였다. 따라서 제안한 주목 연산자를 사용하면 잡음이 많고 복잡한 배경으로부터 물체만을 쉽게 검출할 수 있도록 하였다. 다양한 합성영상(synthetic images)과 실영상(real images)에 대해 실험하여 잡음의 영향을 적게 받으며 효과적으로 다중 물체를 검출함을 확인하였다.

• 대한기계학회논문집A
• /
• 제25권5호
• /
• pp.891-898
• /
• 2001

This paper develops a master manipulator which can reflect a force from a slave manipulator effectively. It many have a big workspace, can represent a human operators manipulation perfectly, and is composed of a position control part, an orientation control part and an end effector control part. This paper also develops a graphic simulator using the Visual C++ and OpenGL in the Window operating system. It can be used to make a virtual slave manipulator and set a virtual working environment, and provide a visual information from a desired view point. A virtual teleoperation system is developed by connecting the developed master manipulator to the graphic simulator using an interfacing hardware bilaterally. It is used for performing a virtual object exploration experiment. In the experiment, two virtual objects are used. They are virtual wall and virtual hexahedron which have 0.7N/mm and 2.2 N/mm stiffness respectively. The experiments are performed under six different working conditions. The experiment results will show the effectiveness of the reflected force from the slave manipulator for improving the efficiency and stability of the teleoperation task.

• 제어로봇시스템학회논문지
• /
• 제22권1호
• /
• pp.8-16
• /
• 2016

In the field of computer vision and robotics, bin picking is an important application area in which object pose estimation is necessary. Different approaches, such as 2D feature tracking and 3D surface reconstruction, have been introduced to estimate the object pose accurately. We propose a new approach where we can use both 2D image features and 3D surface information to identify the target object and estimate its pose accurately. First, we introduce a label detection technique using Maximally Stable Extremal Regions (MSERs) where the label detection results are used to identify the target objects separately. Then, the 2D image features on the detected label areas are utilized to generate 3D surface information. Finally, we calculate the 3D position and the orientation of the target objects using the information of the 3D surface.

• 한국농공학회논문집
• /
• 제54권2호
• /
• pp.115-125
• /
• 2012

Application of the Object-oriented Programming (OOP) method to the Finite Element Model (FEM) program has various strengths including the features of encapsulation, polymorphism and inheritance. However, this technique should be based upon a premise that the independency of the object method and data to be used is guaranteed. By attempting to apply the OOP to the FEM, existing researches go against the independency of the OOP which is an essential feature of the method. The reason is this: existing researches apply the OOP to modules in accordance with analysis procedures, although the data to be used is classified as an element unit in the FEM. Therefore, the required independency cannot be maintained as whole stiffness matrices and boundary conditions are combined together. Also, solutions are sought from analysis module after data is regrouped at the pre-processor, and their results are analyzed during the post-processor. As this is similar to a batch processing, it cannot use data at analysis, and recalculation should be done from the beginning if any condition is changed after the analysis is complete, which are limitations of the existing researches. This research implemented the Object-orientation of elements so that the three features of the OOP (i.e. encapsulation, polymorphism and inheritance) can be guaranteed and their independency maintained as a result. For this purpose, a model called 'Object-oriented Finite element Model ensuring the Independency of Elements (OFMIE)', which enables the analysis of targets through mutual data exchanges within instance, was developed. In conclusion, the required independency was achieved in the instance of the objected elements and the analysis results of previous conditions could be used for the analysis after changes. The number of repetitive calculations was reduced by 75 per cent through this gradual analysis processes.