• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제11권1호
• /
• pp.302-320
• /
• 2017

In this paper, an innovative robust feature detection and matching strategy for visual odometry based on stereo image sequence is proposed. First, a sparse multiscale 2D local invariant feature detection and description algorithm AKAZE is adopted to extract the interest points. A robust feature matching strategy is introduced to match AKAZE descriptors. In order to remove the outliers which are mismatched features or on dynamic objects, an improved random sample consensus outlier rejection scheme is presented. Thus the proposed method can be applied to dynamic environment. Then, geometric constraints are incorporated into the motion estimation without time-consuming 3-dimensional scene reconstruction. Last, an iterated sigma point Kalman Filter is adopted to refine the motion results. The presented ego-motion scheme is applied to benchmark datasets and compared with state-of-the-art approaches with data captured on campus in a considerably cluttered environment, where the superiorities are proved.

• 대한자기공명의과학회:학술대회논문집
• /
• 대한자기공명의과학회 2002년도 제7차 학술대회 초록집
• /
• pp.106-106
• /
• 2002

Purpose： The purposes of our study are (1) to develop a brain phantom which can be used for multimodal image registration, (2) to evaluate the accuracy of image registration with the home-made phantom. Method： A brain phantom which could be used for image registration technique of CT-MR and CT-SPECT images using chamfer matching was developed. The brain phantom was specially designed to obtain imaging dataset of CT, MR, and SPECT. The phantom had an external frame with 4 N-shaped pipes filled with acryl rods for CT, MR imaging and Pb rods for SPECT imaging. 8 acrylic pipes were inserted into the empty space of the brain phantom to be imaged for geometric evaluation of the matching. Accuracy of image fusion was assessed by the comparison between the center points of the section of N-shaped bars in the external frame and the inserted pipes of the phantom. Technique with partially transparent, mixed images using color on gray was used for visual assessment of the image registration process.

• 제어로봇시스템학회논문지
• /
• 제15권9호
• /
• pp.909-915
• /
• 2009

This work is to develop a vision system for high precision alignment between upper and lower plates required at the imprinting process of the large LCD flat panel. We compose a gantry-stage that has highly repeated accuracy for high precision alignment and achieves analysis about thermal transformations of stage itself. Position error in the stage is corrected by feedback control from the analysis. This system can confirm alignment mark of upper and lower plates by using two cameras at a time for the alignment of two plates. Pattern matching that uses geometric feature is proposed to consider the recognition problem for alignment mark of two plates. It is algorithm to correct central point and angle for the alignment from the recognized mark of upper and lower plates based on the special characteristics. At the alignment process, revision for error position is performed through Look and Move techniques.

• 한국건설관리학회논문집
• /
• 제5권5호
• /
• pp.151-162
• /
• 2004

Current methods for construction site modeling employ large, expensive laser range scanners that produce dense range point clouds of a scene from different perspectives. Days of skilled interpretation and of automatic segmentation may be required to convert the clouds to a finished CAD model. The dynamic nature of the construction environment requires that a real-time local area modeling system be capable of handling a rapidly changing and uncertain work environment. However, in practice, large, simple, and reasonably accurate embodying volumes are adequate feedback to an operator who, for instance, is attempting to place materials in the midst of obstacles with an occluded view. For real-time obstacle avoidance and automated equipment control functions, such volumes also facilitate computational tractability. In this research, a human operator's ability to quickly evaluate and associate objects in a scene is exploited. The operator directs a laser range finder mounted on a pan and tilt unit to collect range points on objects throughout the workspace. These groups of points form sparse range point clouds. These sparse clouds are then used to create geometric primitives for visualization and modeling purposes. Experimental results indicate that these models can be created rapidly and with sufficient accuracy for automated obstacle avoidance and equipment control functions.

• 한국측량학회지
• /
• 제37권2호
• /
• pp.55-62
• /
• 2019

The stereo geometry establishment based on the precise sensor modeling is prerequisite for accurate stereo data processing. Ground control points are generally required for the accurate sensor modeling though it is not possible over the area where the accessibility is limited or reference data is not available. For the areas, the relative orientation should be carried out to improve the geometric consistency between the stereo data though it does not improve the absolute positional accuracy. The relative orientation requires conjugate points that are well distributed over the entire image region. Therefore the automatic conjugate point extraction is required because the manual operation is labor-intensive. In this study, we applied the method consisting of the key point extraction, the search space minimization based on the epipolar line, and the rigorous outlier detection based on the RPCs (Rational Polynomial Coefficients) bias compensation modeling. We tested different parameters of window sizes for Kompsat-2 across track stereo data and analyzed the RPCs precision after the bias compensation for the cases whether the epipolar line information is used or not. The experimental results showed that matching outliers were inevitable for the different matching parameterization but they were successfully detected and removed with the rigorous method for sub-pixel level of stereo RPCs precision.

• Asian Pacific Journal of Cancer Prevention
• /
• 제13권5호
• /
• pp.2393-2398
• /
• 2012

Objectives: The purpose of this study was to evaluate the geometric accuracy of thoracic anatomic landmarks as target surrogates of intrapulmonary tumors for manual rigid registration during image-guided radiotherapy (IGRT). Methods: Kilovolt cone-beam computed tomography (CBCT) images acquired during IGRT for 29 lung cancer patients with 33 tumors, including 16 central and 17 peripheral lesions, were analyzed. We selected the "vertebrae", "carina", and "large bronchi" as the candidate surrogates for central targets, and the "vertebrae", "carina", and "ribs" as the candidate surrogates for peripheral lesions. Three to six pairs of small identifiable markers were noted in the tumors for the planning CT and Day 1 CBCT. The accuracy of the candidate surrogates was evaluated by comparing the distances of the corresponding markers after manual rigid matching based on the "tumor" and a particular surrogate. Differences between the surrogates were assessed using 1-way analysis of variance and post hoc least-significant-difference tests. Results: For central targets, the residual errors increased in the following ascending order: "tumor", "bronchi", "carina", and "vertebrae"; there was a significant difference between "tumor" and "vertebrae" (p = 0.010). For peripheral diseases, the residual errors increased in the following ascending order: "tumor", "rib", "vertebrae", and "carina"; There was a significant difference between "tumor" and "carina" (p = 0.005). Conclusions: The "bronchi" and "carina" are the optimal surrogates for central lung targets, while "rib" and "vertebrae" are the optimal surrogates for peripheral lung targets for manual matching of online and planned tumors.

• 대한토목학회논문집
• /
• 제44권2호
• /
• pp.263-270
• /
• 2024

무인항공기와 무인항공기 센서가 다양하게 개발됨에 따라 기존의 항공사진 또는 원격탐사보다 좁은 면적에 대한 정보를 빠르게 업데이트할 수 있다. 하지만 무인항공기 사진측량에서 지상기준점의 획득과 입력은 많은 시간이 소요되며, 지상기준점 측량과 입력이 잘못될 경우 기하 왜곡이 발생한다. 본 연구에서는 이러한 지상기준점 획득과 입력의 시간을 줄이기 위해 RGB 기준 정사영상을 제작하고, 다양한 센서의 목적 정사영상에 특징점 알고리즘을 적용하여 비교·평가를 수행하였다. 연구대상지 2곳에 대해 4가지 특징점 추출 알고리즘을 적용했으며, 그 결과 특징점 대비 매칭쌍의 비율은 speeded up robust features(SURF)가 가장 우수하였다. 전체적으로 비교했을 때 accelerated-KAZE(AKAZE) 방법이 가장 많은 특징점과 매칭쌍을 추출했으며, binary robust invariant scalable keypoints(BRISK) 방법이 가장 적은 특징점과 매칭쌍을 추출했다. 본 결과를 통해 센서별 목적 정사영상 기하보정 수행 시 AKAZE 방법이 우수한 것을 확인할 수 있었다.

• 정보처리학회논문지A
• /
• 제14A권7호
• /
• pp.435-442
• /
• 2007

본 논문에서는 3D 메쉬 모델의 텍스쳐 매핑, 단순화, 모핑, 압축, 형상정합 등 다양한 분야에 응용될 수 있는 메쉬분할 문제를 다룬다. 메쉬 분할은 주어진 메쉬를 서로 떨어진 집합(disjoint sets)으로 나누는 과정으로서, 본 논문에서는 메쉬의 전역적 및 국부적 기하 특성을 동시에 반영하여 메쉬를 분할하는 방법을 제시하고자 한다. 먼저 주어진 메쉬의 국부적 기하 특성인 곡률 정보와 전역적 기하 특성인 볼록성을 이용하여 메쉬 정점들 중 첨예정점(sharp vertex)을 추출하고, 모든 첨예정점들 간의 유클리디언 거리에 기반한 $\kappa$-평균군집화 기법[26]을 적용하여 첨예 정점들을 분할한다. 분할된 첨예정점에 속하지 않는 나머지 정점들에 대해서는 유클리디언 거리상 가까운 군집으로 병합하여 최종적인 메쉬분할이 이루어진다. 또한 본 논문에서 제안한 메쉬분할 방법을 구현하여 여러 메쉬 모델에 대해 실험 결과를 보여준다.

• 한국지리정보학회지
• /
• 제1권1호
• /
• pp.8-17
• /
• 1998

본 논문에서는 지상기준점의 데이터베이스 구축과 영상 GCP의 활용을 소개한다. 고해상도 위성영상 또는 항공사진의 정밀 기하학적 보정을 위해서는 정확한 지상기준점을 추출하는 작업이 필수적이다. 이러한 지상기준점 추출의 편의성과 정확성을 향상시키기 위하여 지상기준점 정보의 데이터베이스를 구축하고 더불어 영상 GCP를 저장하였다. 원시영상의 보정을위해 GCP 추출을 반복할 필요 없이 데이터베이스의 GCP를 이용함으로써 재사용의 이점을 얻는다. 또한 영상 GCP에 대응하는 원시영상의 좌표계를 결정할 수 있는 알고리즘의 구현을 통해서 데이터베이스에 저장된 정확한 GCP들의 값을 다른 원시영상의 보정에 전이시킬 수 있는 효과를 얻을 수 있다.

• 정보보호학회논문지
• /
• 제20권2호
• /
• pp.15-21
• /
• 2010

지문정보를 이용한 사용자 인증은 편리함과 동시에 강력한 보안을 제공한다. 그러나 유한개의 손가락 개수로 인해 패스워드처럼 변경을 자유롭게 할 수 없기 때문에 사용자 인증을 위해 저장된 지문정보가 타인에게 도용된다면 심각한 문제를 일으키게 된다. 지문정보를 안전하게 보호하기 위한 지문 퍼지볼트 시스템에 대한 연구가 활발히 진행되고 있다. 또한, 기하학적 해싱 기법을 적용하여 지문 정렬 문제를 해결하려는 연구도 진행되었다. 본 논문에서는 기하학적 해싱 기반 지문 퍼지볼트 시스템의 하드웨어 구조를 제안한다. 제안한 구조는 소프트웨어 모듈과 하드웨어 모듈로 통합 구성된다. 하드웨어 모듈은 등록 해시 테이블과 인증 해시 테이블의 변환 특징점의 정합을 담당한다. 실험결과, 제안한 하드웨어 구조의 수행 시간은 지문 특징점의 수가 36개이고 거짓 특징점의 수가 100개 일 때 0.2초, 400개 일 때, 0.53초이다.