• Journal of the Korean Association of Geographic Information Studies
• /
• v.20 no.1
• /
• pp.152-161
• /
• 2017

Making precise 3D maps using Mobile Mapping System (MMS) sensors are essential for the development of self-driving cars. This paper conducts research on the extraction of 3D objects around the roads using the point cloud acquired by the MMS Light Detection and Ranging (LiDAR) sensor through the following steps. First, the digital surface model (DSM) is generated using MMS LiDAR data, and then the slope map is generated from the DSM. Next, the 3D objects around the roads are identified using the slope information. Finally, 97% of the 3D objects around the roads are extracted using the morphological filtering technique. This research contributes a plan for the application of automated driving technology by extracting the 3D objects around the roads using spatial information data acquired by the MMS sensor.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2011.11a
• /
• pp.365-368
• /
• 2011

본 논문은 계속적으로 움직이는 Pan & Tilt 카메라를 가지고 객체를 안정적으로 추적하기 위해 Level Set 알고리즘과 Pan & Tilt 카메라간의 상호 유동적인 시스템을 설계하는 방법에 대하여 기술하고 있다. 특정객체를 Pan & Tilt 카메라로 계속적으로 추적하고자 할 때는 안정적인 배경영상을 얻을 수 없기 때문에 MOG 와 같은 통계적인 추적알고리즘을 쓰는 것이 불가능해진다. 본 논문에서는 배경 영상이 계속적으로 변하기 때문에 고정된 배경 영상을 가질 수 없는 문제와 이로 인해 객체의 영역을 잘 추출할 수 없다는 한계를 극복하기 위해 Level Set 에 기반한 외곽선 추적 방법을 이용한다. 이 방법은 단지 차영상만을 가지고도 어느 정도 객체의 영역을 추출할 수 있는 방법이다. Level Set 방법은 높은 복잡도를 가지기 때문에 실시간 계산이 빠른 외곽선 추적 방법을 이용하였으며, 이를 통해 실시간 영상에 대한 외곽선 추적을 가능하게 하였다. 그리고 Level Set 기반 외곽선 추출 방법에 의해 객체의 중심점을 구하는 알고리즘과 Pan & Tilt 카메라에 의해 객체를 추적하는 알고리즘 간에 유동적인 연결을 하였다.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.15 no.3
• /
• pp.579-586
• /
• 2020

There is a large amount of multimedia data on the web page, and a method of extracting semantic information from low level visual information for accurate retrieval is being studied. However, most of these techniques extract one of information from a single image, so it is difficult to extract semantic information when multiple objects are combined in the image. In this paper, each low-level feature is extracted to extract various objects and backgrounds in an image, and these are divided into predefined backgrounds and objects using SVM. The objects and backgrounds divided in this way are constructed with ontology, infer the semantic information of location and association using inference engine. It's possible to extract the semantic information. We propose this method process the complex and high-level semantic information in image.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.12 no.4
• /
• pp.1897-1903
• /
• 2011

This paper discusses a 2D/3D conversion of images using technologies like object extraction and depth-map creation. The general procedure for converting 2D images into a 3D image is extracting objects from 2D image, recognizing the distance of each points, generating the 3D image and correcting the image to generate with less noise. This paper proposes modified new methods creating a depth-map from 2D image and recognizing the distance of objects in it. Depth-map information which determines the distance of objects is the key data creating a 3D image from 2D images. To get more accurate depth-map data, noise filtering is applied to the optical flow. With the proposed method, better depth-map information is calculated and better 3D image is constructed.

• The Transactions of the Korea Information Processing Society
• /
• v.6 no.4
• /
• pp.941-951
• /
• 1999

Class components in a repository require exact information representation by reason of insufficiency of various visual information through search and extraction. In this paper, we have described syntax-analysis method and viewer facilities about class components. In the components analysis, we extracted class information which consists of class_names, methods, attributes, class inheritance relationship, and graphic information. viewer represents extracted class information and creates a new class. Also, it provides facilities of reuse, insertion, and deletion. Not only Viewer represents class hierarchy diagram, and shows detail information about each class, but also it creates new classes. In this paper, we implemented a Viewer using the components, inheritance, diagram information and process. And we enhanced understanding of programs through viewer, and supported prototype for class creation.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2004.05a
• /
• pp.245-248
• /
• 2004

본 연구는 레거시 시스템의 인터페이스 정보로부터 의미 있는 정보를 파악하여 새로운 시스템에 통합될 수 있도록 하기 위한 기존 레거시 시스템의 인터페이스에 기반한 객체 추출 기법(이하 TEILOR ; Techniques to extract Object based on Interface of Legacy System for Object Reusability)을 제안한다. 본 논문에서 제안하는 TEILOR는 인터페이스 사용사례 분석 단계, 인터페이스 객체 분할 단계, 객체구조 모델링 단계, 객체 모델 통합 단계 등 4단계로 구성되어 있다. 인터페이스 사용사례 분석 단계는 인터페이스 구조, 레거시 시스템과 사용자간의 상호작용 정보를 획득하는 단계이다. 인터페이스 객체분할 단계는 인터페이스 정보를 의미 있는 필드들로 구분하는 단계이며, 객체구조 모델링 단계는 인터페이스 객체들간의 구조적 관계와 협력 관계를 파악하여 모델링하는 단계이다. 마지막으로 객체 모델 통합 단계는 객체 단위의 단위 모델들을 통합하여 추상화된 정보를 포함한 상위 수준의 통합 모델을 유도하는 단계다. TEILOR에 의해 생성된 객체 통합 모델은 역공학 기술자들의 레거시 시스템 이해와 레거시 시스템의 정보를 새로운 시스템에 적용하는데 있어 효율성을 극대화할 수 있다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2016.06a
• /
• pp.273-274
• /
• 2016

본 논문에서는 다중 랜덤 워커(multiple random walkers)에 기반한 객체 추적 기법을 제안한다. 우선 서포트 벡터 머신(support vector machine)을 이용한 분류기 기반 객체 추적 기법을 소개한다. 다음으로 영상의 영역에 대한 특징 벡터 중 배경으로부터 추출된 특징 벡터를 억제하는 기법을 제안한다. 영역에서 배경 요소를 찾기 위해 다중 랜덤 워커를 이용한 전경 및 배경 추출 방법을 제시한다. 배경 요소를 억제하여 학습된 서포트 벡터 머신은 객체와 배경이 유사한 영상, 객체가 다른 물체에 의해 가려지는 영상 등에서 객체와 배경을 확실하게 구분하여 객체를 잃지 않고 추적할 수 있다. 마지막으로 실험을 통해 제안하는 기법이 기존 기법에 비해 우수한 추적 성능을 보임을 확인한다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.1C
• /
• pp.92-101
• /
• 2004

This paper proposes a algorithm to extract the variety region (object) from video for the real-time encoding of MPEG-4 based. The previous object segmentation methods cannot used the videophone or videoconference required by real-time processing. It is difficult to transfer a video to real-time because it increased complexity for the operation of each pixel on the spatial segmentation and temporal segmentation method proposed by MPEG-4 Working Group. But algorithm proposed for this thesis not operates a pixel unit but operates a macro block unit. Thus this enables real-time transfer. But this algorithm cannot extract several object for a image using proposed algorithm as previous algorithm. On system constructed by encoder and decoder. A proposed algorithm inserted for encoder as pre-process.

• The KIPS Transactions:PartB
• /
• v.15B no.6
• /
• pp.501-506
• /
• 2008

The classical snake algorithm has a problem in detecting the boundary of an object with deep concavities. While the GVF method can successfully detect boundary concavities, it consumes a lot of time computing the energy map. In this paper, we propose an algorithm to reduce the computation time and improve performance in detecting the boundary of an object with high concavity. We define the degree of complexity of object boundary as the local curvature. If the value of the local curvature is greater than a threshold value, new snake points are added. Simulation results on several different test images show that our method performs well in detecting object boundary and requires less computation time.

• Journal of Convergence for Information Technology
• /
• v.7 no.6
• /
• pp.229-235
• /
• 2017

In recent years, research has been actively carried out to recognize and recognize objects based on a large amount of data. In this paper, we propose a system that extracts objects that are thought to be obstacles in road driving images and recognizes them by car, man, and motorcycle. The objects were extracted using Optical Flow in consideration of the direction and size of the moving objects. The extracted objects were recognized using Alexnet, one of CNN (Convolutional Neural Network) recognition models. For the experiment, various images on the road were collected and experimented with black box. The result of the experiment showed that the object extraction accuracy was 92% and the object recognition accuracy was 96%.