자기 일치성을 이용한 다중 영상 스테레오 기법

우동민,김민석 명지대학교 산업기술연구소 2001 産業技術硏究所論文集 Vol.20 No.-

The ability to efficiently and robustly recover accurate 3D terrain models from sets of stereoscopic images is important to many civilian and military applications. To develop an effective and practical terrain modeling system, we propose a new multi-image stereo method which detect unreliable elevations in DEM(digital elevation map), and fuse several DEMs from multiple sources into an accurate and reliable result. This paper focuses on two key factors for generating robust 3D terrain models: the ability to detect unreliable elevation estimates and the ability to fuse the reliable elevations into a single optimal terrain model. We apply the self-consstency methodology to reconstruct accurate DEM from multi-image and show the method is more effective than the conventional stereo image 3D reconstruction method. Using photo-realistic simulator, four synthetic images are generated from ground truth DEM and orthoimage to evaluate the accuracy of the proposed method quantitatively.

위성 영상을 이용한 3차원 영상처리

우동민 한국공학기술학회 2002 공학교육과 기술 Vol.9 No.3

신경망을 이용한 렌즈 왜곡 모델을 이용한 카메라 보정

우동민,김민석 명지대학교 산업기술연구소 2001 産業技術硏究所論文集 Vol.20 No.-

A new camera calibration method based on a lens distortion model by an artificial neural network is proposed for three dimensional computer vision problems. The camera calibration procedure using the proposed method employs the widely used two step method. In the first step, the distortion-free linear camera model is constructed by estimating parameters of the model. In the second step, the proposed neural network model is established for mapping the distorted image points with the ideal image points. In order to show the performance of the proposed method, images from two different cameras with two different camera angles were used for calibrating the cameras. The performance of the proposed neural network approach is compared with the well-known Tsai's two stage method in terms of calibration errors. The results show that the proposed approach gives much more stable and acceptable calibration error over Tsai's two stage method regardless of camera resolution and camera angles.

스테레오 정합을 이용한 3차원 재구성 과정의 정량적 평가

우동민,Woo, Dong-Min 한국전기전자학회 2013 전기전자학회논문지 Vol.17 No.1

3차원 영상 해석 기법에 의해 구해진 DEM(Digital Elevation Map)을 정량적으로 평가하는 것은 영상 해석 기법의 유효성을 검증하기 위해 매우 중요하다. 본 논문에서는 모의 영상 제작에 의한 3차원 재구성 과정의 새로운 정량적 평가 방법을 제안하였다. 제안된 방법은 미리 확보된 DEM과 정사영상(ortho-image)을 가상의 실제 값(pseudo ground truth)으로 가정한 것에 의한 것이다. 제안된 방법의 과정은 그래픽스에서 사용되는 ray tracing 알고리즘을 구성하여 가상의 실제 값에 적용함으로서 원하는 시점으로부터의 한 쌍의 모의 영상을 제작하는 것으로부터 시작된다. 제작된 모의 영상 쌍으로부터 구해진 DEM을 가상의 실제 값과 비교하면 구해진 DEM의 정량적인 오차 분석이 가능하여, 적용된 3차원 영상 해석 기법의 유효성이 평가될 수 있다. 제안된 평가 방법의 타당성을 검증하기 위해, 정량적 및 정성적인 실험이 수행되었다. 이를 위해 발생되는 모의 영상이 실제 형상을 재현하는 정도를 정량적인 수치로서 구하여 제안된 방법을 타당성을 입증하였다. 또한 정합창의 크기 변화에 따른 DEM의 정확도를 제안된 평가 방법에 의해 실험하였다. 이러한 실험 결과가 예견된 결과와 일치함에 의해 제안된 평가 방법의 타당성을 정성적으로도 명백히 증명하였다. The quantitative evaluation of DEM(Digital Elevation Map) is very important to the assessment of the effectiveness for the applied 3D image analysis technique. This paper presents a new quantitative evaluation method of 3D reconstruction process by using synthetic images. The proposed method is based on the assumption that a preacquired DEM and ortho-image should be the pseudo ground truth. The proposed evaluation process begins by generating a pair of photo-realistic synthetic images of the terrain from any viewpoint in terms of application of the constructed ray tracing algorithm to the pseudo ground truth. By comparing the DEM obtained by a pair of photo-realistic synthetic images with the assumed pseudo ground truth, we can analyze the quantitative error in DEM and evaluate the effectiveness of the applied 3D analysis method. To verify the effectiveness of the proposed evaluation method, we carry out the quantitative and the qualitative experiments. For the quantitative experiment, we prove the accuracy of the photo-realistic synthetic image. Also, the proposed evaluation method is experimented on the 3D reconstruction with regards to the change of the matching window. Based on the fact that the experimental result agrees with the anticipation, we can qualitatively manifest the effectiveness of the proposed evaluation method.

3D Line Segment Extraction Based on Line Fitting of Elevation Data

우동민 한국전기전자학회 2009 전기전자학회논문지 Vol.13 No.2

In this paper, we are concerned with a 3D line segment extraction method by area-based stereo matching technique. The main idea is based on line fitting of elevation data on 2D line coordinates of ortho-image. Elevation data and ortho-image can be obtained by well-known area-based stereo matching technique. In order to use elevation in line fitting, the elevation itself should be reliable. To measure the reliability of elevation, in this paper, we employ the concept of self-consistency. We test the effectiveness of the proposed method with a quantitative accuracy analysis using synthetic images generated from Avenches data set of Ascona aerial images. Experimental results indicate that our method generates 3D line segments almost 7.5 times more accurate than raw elevations obtained by area-based method.

DEM 융합에 의한 3차원 지형 복원

우동민 明知大學校 産業技術硏究所 2004 産業技術硏究所論文集 Vol.23 No.-

The ability to efficiently and robustly recover accurate 3D terrain models from sets of stereoscopic images is important to many civilian and military applications. A stereo matching has been an important tool for reconstructing three dimensional terrain. However, there exit many factors causing stereo matching error, such as occlusion, no feature or repetitive pattern in the correlation window, intensity variation, etc. Among them, occlusion can be only resolved by true multi-image stereo. In this paper, we present multi-image stereo method using DEM fusion as one of efficient and reliable true multi-image methods. Elevations generated by all pairs of images are combined by the fusion process which accepts an accurate elevation and rejects an outlier. We propose three fusion schemes: THD(Thresholding), BPS(Best Pair Selection) and MS(Median Selection). THD averages elevations after rejecting outliers by thresholding, while BPS selects the most reliable elevation. To determine the reliability of a elevation or detect the outlier, we employ the measure of self-consistency. The last scheme, MS, selects the median value of elevations. We test the effectiveness of the proposed methods with a quantitative analysis using simulated images. Experimental results indicate that all three fusion schemes showed much better improvement over the conventional binocular stereo in natural terrain of 29 Palms and urban site of Avenches.

중심신경망을 이용한 3차원 선소의 군집화에 의한 위성영상의 3차원 건물모델 재구성

우동민,박동철,Hai Nguyen Ho,김태현 大韓遠隔探査學會 2011 大韓遠隔探査學會誌 Vol.27 No.2

본 논문에서는 중심신경망을 이용하여 위성영상으로부터 직사각형 형태의 3차원 건물의 지붕모델을 재구성하는 방법을 연구하였다. 제안된 3차원 지붕모델 재구성 기법의 핵심은 3차원 선소의 군집화에 있다. 이를 위해 한 쌍의 스테레오 영상으로부터 구해진 DEM (Digital Elevation Map) 데이터와 2차원 선소에 의해서 3차원 선소를 발생하였다. 제안된 군집화 과정은 중심신경망을 이용한 방법에 의해 수행되며, 2단계로 구성된다. 첫 번째 단계에서는 선소 추출과정에서 끊어지거나, 중복된 3차원 선소를 건물을 이루는 주된 선소로 군집화하고, 두 번째 단계에서는 건물을 구성하는 주된 선소를 구하기 위해 서로 평행인 선소들의 군으로 군집화를 수행한다. 이 군집화 결과를 최종 클러스터링 과정을 통해 직사각형 형태의 지붕모델로 재구성하게 된다. 제안된 방법이 대전지역의 고해상도 IKONOS 위성영상에 의해 실험되었다. 재구성된 건물모델이 원래 건물의 위치와 형태를 대체로 정확히 반영하여, 본 논문에서 제안된 기법을 고해상도 위성영상에 적용하여 도시지역의 건물모델을 구축하는데 효과적으로 사용될 수 있음이 입증되었다. This paper highlights the reconstruction of the rectilinear type of 3D rooftop model from satellite image data using centroid neural network. The main idea of the proposed 3D reconstruction method is based on the grouping of 3D line segments. 3D lines are extracted by 2D lines and DEM (Digital Elevation Map) data evaluated from a pair of stereo images. Our grouping process consists of two steps. We carry out the first grouping process to group fragmented or duplicated 3D lines into the principal 3D lines, which can be used to construct the rooftop model, and construct the groups of lines that are parallel each other in the second step. From the grouping result, 3D rooftop models are reconstructed by the final clustering process. High-resolution IKONOS images are utilized for the experiments. The experimental results indicate that the reconstructed building models almost reflect the actual position and shape of buildings in a precise manner, and that the proposed approach can be efficiently applied to building reconstruction problem from high-resolution satellite images of an urban area.

적응적인 스테레오 정합 기법에 의한 3차원 지형 복원

禹東玟(Dong-Min Woo) 대한전기학회 1998 전기학회논문지 Vol.47 No.12

신경망을 이용한 렌즈의 왜곡모델 구성 및 카메라 보정

김민석,우동민 明知大學校 産業技術硏究所 2000 産業技術硏究所論文集 Vol.19 No.-

The objective of camera calibration is to determine the internal optical characteristics of camera and 3D position and orientation of camera with respect to the real world. Generally, camera calibration should be. applicable to general purpose cameras and lenses. The general method to improve the accuracy of the calibration is by using mathematical distortion of lens. In this paper. we propose the method which calibration distortion model by using neural network. The neural network model implicitly reflects all the distortion model. We can predict the high accuracy of calibration method proposed in this paper. Neural network can set properly the distortion model which has difficulty to estimate exactly in general method. The performance of the proposed neural network approach is compared with the well-known Tsai's two stage method in terms of calibration errors. The results show that the proposed approach gives more stable and acceptable improvement over Tsai's two stage method regardless of camera resolution and camera angle.

시변 망각 인자를 사용한 최소 자승 추정의 분산 감소 극점 배치 자기 동조 알고리즘에 관한 연구

이상배,박민용,민용기,박찬영,우동민 연세대학교 산업기술연구소 1990 논문집 Vol.22 No.1

In this paper, the explicit self-tuning pole assignment controller with variable forgetting factors is generalized to allow the output and/or input variance to be reduced. The Algorithm can give significant reductions in variance for little extra computational effort and is presented for servotracking using least-squres estimation. Moreover, the use of a variable forgetting factor with correct choice of information bound can avoid 'blowing-up' of the covariance martix of the estimates and subsequent unstable control. At each step a weighting factor is chosen to maintain constant scalar measure of the information content of the estimator. Such as approach enables the parameter estimates to follow both slow and sudden changes in the plant dynamics.