http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Lee, Dong-Cheon,Lee, David H.,Lee, Dae Geon Hindawi Limited 2019 Journal of sensors Vol.2019 No.-
<P>Light detection and ranging (LiDAR) data collected from airborne laser scanner system is one of the major sources to reconstruct Earth’s surface features. This paper presents a method for detecting model key points (MKPs) of the buildings using LiDAR point clouds. The proposed approach utilizes shaded relief images (SRIs) derived from the LiDAR data. The SRIs based on the concept of the shape from shading could provide unique information about individual surface patches of the building roofs. The main advantage of the proposed approach is to detect directly MKPs, which are primitives for 3D building modeling, without segmenting point clouds. Depending on the location of the light source, the SRIs are created differently. Therefore, integration of the multidirectional SRIs created from different locations of the light source could provide more reliable results. In addition, the vertical exaggeration (i.e., scaling<I> Z</I>-coordinates) is also beneficial because constituent surface patches of the roofs in the SRIs created with vertically exaggerated LiDAR data are more distinguishable. To determine the MKPs of the roofs, building data was separated from other objects using modified marker-controlled watershed algorithm in accordance with criteria to specify buildings such as area, height, and standard deviation. This process could remove the unnecessary objects such as trees, vegetation, and cars. The curvature scale space (CSS) corner detector was used to determine MKP since this method is robust to geometric changes such as rotation, translation, and scale. The proposed method was applied to simulated and real LiDAR datasets with various roof types. The experimental results show that the proposed method is effective in determining MKPs of various roof types with high level of detail (LoD).</P>
이대건(Lee, Dae Geon),이동천(Lee, Dong-Cheon) 한국측량학회 2020 한국측량학회 학술대회자료집 Vol.2020 No.7
Object recognition and classification are important in high-level image processing such as computer vision and machine learning. Detection, recognition, identification, and classification are sequential and progressive learning procedures. However, implementation of human-like learning mechanism using artificial neural network (ANN) with limited training data is challenging task. This paper proposes deep learning for land cover classification using shaded relief maps created from DSM as training data sets. The results show that the derived feature information from original data increases number of training data and improves performance of ANN.
라이다 데이터로부터 생성된 음영기복도를 이용한 특징점 검출
이대건(Lee, Dae Geon),이동천(Lee, Dong-Cheon) 한국측량학회 2017 한국측량학회 학술대회자료집 Vol.2017 No.4
During last decades, many efforts have been made to automation of the object modeling using LiDAR data. Most of the object modeling methods require segmenting surface patches. This paper aims to determine MKPs (Model Key Points) of the objects from SRM (Shaded Relief Map) generated by LiDAR data. The data lacks visual information, but provide 3D coordinates of the object surfaces explicitly. SRM has no geometric distortion unlike aerial images. In addition, shading is an important element of shape-from-X that can recognize object shape in CV (Computer Vision). Corner detector was used to determine MKPs in SRMs that generated from multiple directions of the light source because corner detection from SRM generated with single direction could not be completely.
영상과 수치지형도의 정합에 의한 위성영상의 Geocoding
이대건(Lee, Dae Geon),이동천(Lee, Dong-Cheon) 한국측량학회 2016 한국측량학회 학술대회자료집 Vol.2016 No.4
Recently, more various kinds of image could be obtained from sensors in different types of platform. Images provide visual information especially geometric characteristics including shape, size and pattern of the features. In order to improve utilization of the images, it should be able to obtain absolute position information from images through geocoding process. In this paper, matching between satellite image and digital topographic map was peformed to obtain control features both on the image and map simultaneous. Matching should be scale, rotation and shift invariant. In these regards, SURF matching method was applied. The corresponding points on image and map from matching were used as control points to compute coordinate transformation parameters. The result was compared with manual measurement and accuracy of the geocoding was evaluated.
이대건(Lee, Dae Geon),김국희(Kim, Guk Hee),이동천(Lee, Dong-Cheon) 한국측량학회 2017 한국측량학회 학술대회자료집 Vol.2017 No.4
Airborne LIDAR system (ALS) directly collects 3D coordinates of the surface. Multiple flight paths are required to acquire data over the large area with overlap areas between strips. Data in the overlap areas has discrepancy due to accuracy of the laser scanner, GPS/INS, calibration, and flying condition. Strip adjustment is necessary to reduce such discrepancy between strips. In this paper, transformation parameters in overlapping strip were computed using conjugate points that were determined by SURF (Speeded Up Robust Features). Finally, accuracy of the strip adjustment parameters was evaluated and RMSE (Root Mean Square Error) of the overlapping areas data after strip adjustment were calculated.
이대건(Lee, Dae Geon),이동천(Lee, Dong-Cheon) 한국측량학회 2016 한국측량학회 학술대회자료집 Vol.2016 No.4
Object modeling is one of the important tasks for establishing spatial information infrastructure. During last several decades, many efforts have been made to automation of the object modeling. The major sources of spatial data are optical images and LiDAR data from different platforms. Since LiDAR data provide 3D coordinates of the object surfaces explicitly, more intrinsic information about objects could be obtained compared with images. This paper aims to determine model key points (MKPs) of the objects from point clouds provided by LiDAR data. While most of the object modeling methods require segmenting surface patches, the proposed method could extract MKPs directly from point cloud data and makes possible to analyze shape of the objects.
이대건(Lee, Dae Geon),이동천(Lee, Dong-Cheon) 한국측량학회 2018 한국측량학회 학술대회자료집 Vol.2018 No.4
항공 레이저 시스템 (ALS)으로부터 획득한 라이다 데이터를 이용하여 3차원 건물 모델링을 위하여 건물데이터를 추출하고 분리하는 과정이 요구된다. 본 연구에서는 라이다 데이터로부터 건물을 분리하기 위하여 향상된 watershed 기법인 marker-controlled watershed를 적용하고, 분리된 데이터로부터 건물의 특성을 고려하여 건물/비건물을 분류하였다. 최종적으로 수목 및 구조물 등 비건물 객체를 제거하고, 건물 데이터를 개별적으로 분리하는 방법을 제안하였다.