• 한국BIM학회 논문집
• /
• 제11권4호
• /
• pp.42-52
• /
• 2021

Paneling building facades is one of the essential procedures in building construction. Traditionally, it has been an easy task of simply projecting paneling patterns drawn in drawing boards onto 3d building facades. However, as many organic or curved building shapes are designed and constructed in modern architectural practices, the traditional one-to-one projection is becoming obsolete for the building types of the kind. That is primarily because of the geometrical discrepancies between 2d drawing boards and 3d curved building surfaces. In addition, curved compound surfaces are often utilized to accommodate the complicated spatial programs, building codes, and zoning regulations or to achieve harmonious geometrical relationships with neighboring buildings in highly developed urban contexts. The use of the compound surface apparently makes the traditional paneling pattern projection more challenging. Various mapping technics have been introduced to deal with the inabilities of the projection methods for curved facades. The mapping methods translate geometries on a 2d surface into a 3d building façade at the same topological locations rather than relying on Euclidean or Affine projection. However, due to the intrinsic differences of the planar 2d and curved 3d surfaces, the mapping often comes with noticeable distortions of the paneling patterns. Thus, this paper proposes a practical method of drawing paneling patterns directly on a curved compound surface utilizing Geodesic, which is faithful to any curved surface, to minimize unnecessary distortions.

• Journal of the Optical Society of Korea
• /
• 제18권6호
• /
• pp.672-678
• /
• 2014

We propose a method to estimate the 3-D shape of surfaces with specular reflection, using a model of the difference in appearance between images reflected from a flat surface and a curved surface. First, we analyze the geometry of spatial reflection from a specular surface and how reflected light varies due to a curved surface. This is used to estimate 3-D shape. The proposed method is shown to be effective in experiments using illumination from spatially distributed light sources and a camera capturing the reflected light from curved, specular surfaces.

• 한국의류산업학회지
• /
• 제19권5호
• /
• pp.635-645
• /
• 2017

The industry of character toys is increasing and new characters are constantly being developed. However, the development of 2D cover patterns for toys is time-consuming due to frequent pattern modifications made through trial and error. Studies are now underway to obtain 2D clothing patterns from 3D body data, however, little research has been done on 2D pattern of character toys. This study suggests efficient guidelines to develop 2D cover patterns with a reasonable accuracy and processing time. Two 3D models of a dog and rabbit were used to develop 2D cover patterns. Independent variables of this study are set as 3 levels of triangle area (small, medium, and large) that influence the efficacy of 3D and 2D pattern development. The determination of the appropriate triangular area was based on the area and shape change of the 2D pattern. A medium or large triangle area was shown to be suitable for a character dog with a smooth curved surface. However, the appropriate triangle area was small if the characteristics of the curved surface are complicated as in the case of rabbit. The head of a dog (a double-curved surface) and the curved forepaw of a rabbit (a triangular area) should be small when the characteristics of the curved surface (such as the hind leg of a rabbit having a large convex surface and a small surface area) are complicated. Grouping by 3D surface characteristics could be a suitable guideline for the triangle area selection.

• 에너지공학
• /
• 제28권1호
• /
• pp.57-64
• /
• 2019

본 연구에서는 산업용 발전설비 부품의 곡면을 표면처리 하기 위한 브라스팅 노즐을 설계한 후 압축성 유동해석을 통해 분사 성능을 살펴보았다. 이때 곡면 노즐의 출구는 부품의 표면을 따라 곡면으로 설계되었다. 3차원 프린팅으로 노즐을 제작한 후 실린더형 시편의 표면에 연마재를 분사하여 세정 성능시험을 수행하였다. 해석 후 얻은 유효 세정 면적과 실험 후 얻은 유효 세정 면적의 크기와 형상이 유사하였으며 이로부터 곡면형 노즐 설계의 타당성 및 실효성을 확인하였다.

• 한국정보과학회논문지:시스템및이론
• /
• 제34권1호
• /
• pp.38-48
• /
• 2007

본 논문에서는 NURBS(Non-Uniform Rational B-Splines)로 표현된 3차원 곡면을 복원하기 위한 영상 기반 곡면 모델링 기법을 제안한다. 카메라 보정이 수행된 복수의 영상으로부터 사용자가 대응곡선을 지정하면 카메라의 정보를 이용하여 대응 곡선의 3차원 복원을 수행한다. 사용자 인터페이스에 의해 각 영상에서 곡선을 지정하기 위하여 본 논문에서는 NURBS 곡선을 이용하였으며 bilinear surface, ruled surface, generalized cylinder, surface of revolution등의 기본 곡면뿐 아니라 skinned surface, swept surface, boundary patch등의 고급 곡면 생성을 지원하여 다양한 모양의 곡면 물체를 모델링 할 수 있다. 또한 영상을 기반으로 view-dependent 텍스처 생성방법을 지원하여 좀더 사실적인 곡면의 복원을 수행한다. 생성된 3차원 곡선 또는 곡면은 VRML과 같은 표준 형식으로 출력하여 재사용할 수 있다.

• 전자공학회논문지SC
• /
• 제49권4호
• /
• pp.1-9
• /
• 2012

본 논문은 금속이나 유리와 같이 정반사 특성을 갖는 경면반사체의 3차원 형상 추정을 위한 영상처리에 관한 것으로, 조명계와 CCD 카메라 및 경면반사체로 이루어진 머신비전 시스템의 기하학적 광로분석을 통해 반사체의 형상이 반사광의 광로에 주는 영향을 수식화하고, 이를 반사체의 3차원 형상 추정에 이용하였다. 즉, 공간적으로 분포된 점광원들이 반사체에 정반사되어 취득된 영상에서, 반사체의 형상에 따라 영상에서의 점광원의 상대적 위치가 변화되므로, 이 변화량을 이용하여 3차원 정보를 추정하는 방법을 제시하고, 실험을 통하여 반사체의 3차원 형상의 추정이 가능함을 보인다.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2005년도 춘계학술대회 논문집
• /
• pp.749-752
• /
• 2005

In the industry, three-dimensional coloring has been needed for realistic prototype from rapid prototyping. Z-corporation developed a 3D printer which provides three-dimensional colored prototype. However, the existing process cannot be adopted to models from other rapid prototyping process. In addition, time and cost for manufacturing colored prototype still remain to be improved. In this study, a new coloring process using ink-jet head is proposed for color printing on three-dimensional prototype surface. Process parameters such as the angle and the distance between ink-jet nozzle and the three-dimensional surface should be investigated from experiments. The correction matrix according to sloped angle to minimize the distortion of 2D image was proposed by analysis of printing error. Therefore, approximated method for angle and discrete length according to the radius of curvature for printing on the curved surface was proposed. By printing image on the doubly curved surface, the method was verified. As a practical example, helmet was chosen for printing images on the curved surface. The character images were applied with approximated method for angle and discrete length and was printed on the helmet surface.

• 패션비즈니스
• /
• 제20권4호
• /
• pp.153-171
• /
• 2016

With the advent of smart clothing for health care and sports, the sophisticated designs with curved seams are drawing attention. One of the problems in those clothing is to determine the design curves in 2D pattern, such that it corresponds to the lines on the intended 3D body. Moreover, the difficulty increases when the original pattern needs to be changed for various sizes and body types. We compare two methods of pattern enlargement in this paper: one is the offset/projection type, and the other is the split grading type. For the enlarged pattern with offset/projection type, the 3D surface offset was first adopted to transform the standard lower body to the target larger size; next, the design lines were projected to the new 3D surface, following which the 3D pattern was developed from the newly transformed 3D surface. In the second method, the enlarged pant patterns were developed by the split grading method. Here, a 3D pattern was developed from the initial body, and then enlarged to the target size by the conventional split grading method. Two feminine pants patterns were examined by 3D virtual fitting. We observed that the 3D offset/projection pants pattern was well fitted, having an evenly distributed surplus, as compared with the sample developed using the split grading method. The difference between the two patterns were apparent at the location where several curved lines merged.

• 한국정밀공학회:학술대회논문집
• /
• 한국정밀공학회 2006년도 추계학술대회 논문집
• /
• pp.129-130
• /
• 2006

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1992년도 한국자동제어학술회의논문집(국제학술편); KOEX, Seoul; 19-21 Oct. 1992
• /
• pp.399-404
• /
• 1992

Curved 3D objects represented by range data contain large amounts of information compared with planar objects, but do not have distinct features for matching to those of object models. This makes it difficult to represent and identify a general 3D curved object. This paper introduces a new approach to representing and finding a holdsite of general 3D curved objects using range data. We develop a three-dimensional generalized Hough transformation which can be also applied to general 3D curved object recognition and which reduces both the computation time and storage requirements. Our approach makes use of the relative geometric differences between particular points on the object surface and some model points which are prespecified arbitrarily and task dependently.