• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.26 no.6
• /
• pp.633-641
• /
• 2008

This study introduces the method of detecting and restoring occlusion areas by using epipolar algorithm and K-means classification algorithm for true ortho-photo generation. In the past, the techniques of detecting occlusion areas are using the reference images or information of buildings. But, in this study the occlusion areas can be automatically detected by using DTM data and exterior orientation parameters. The detected occlusion areas can be restored by using anther images or the computed values which are determined in K-means classification algorithm. In addition, this method takes advantages of applying epipolar algorithm in order to find same location in overlapping areas among images.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.28 no.1
• /
• pp.83-92
• /
• 2010

Emergence of high-resolution digital aerial cameras and airborne laser scanners have made innovative progress in photogrammetry and spatial information technology. The purpose of this study is to generate true orthoimage by recovering occlusion areas. The orthoimages were generated patch-based transformation. The occlusion areas were mutually corrected by using multiple aerial images. This study proposed a novel method of building roof based orthoimage generation and an effective method of occlusion area detection and recovery. The proposed methods could be efficient to generate true orthoimages in urban areas where occlusion areas are problematic.

• Proceedings of the KSRS Conference
• /
• 2000.04a
• /
• pp.143-148
• /
• 2000

레이저 프로파일링 시스템의 등장으로, 기존에는 얻을 수 없었던 도시 지역에 대한 DTM 취득이 가능해졌고, 더욱 정확한 정사투영영상 또한 제작할 수 있게 되었다. 하지만, 높이 변화를 보이는 자연지물과 인공구조물이 있는 지역에 대해 기존의 정사투영사진 제작기법이 적용될 때, 폐색이나 이중매핑과 같은 문제가 발생하게 된다. 특히 고층건물이 밀집되어 있는 도심지에서 이러한 현상은 두드러져 정사투영영상의 품질을 저해하는 주요한 원인이 된다. 따라서, 본 연구에서는 카메라의 외부표정요소와 DTM을 이용하여 폐색영역을 탐지하고, 폐색이 안된 다른 영상의 정보를 통해 폐색영역을 복원하여 더욱 완전한 정사투영을 제작할 수 있는 알고리즘을 제안하였다. 제안된 알고리즘에 의해 자연지물이나 인공고조물에 의한 폐색영역을 탐지할 수 있었고 폐색영역의 많은 부분을 부가영상을 이용하여 복원하였다. 건물에 대한 사전지식을 이용하여 폐색영역을 탐지하는 국내 연구가 있지만, 본 연구는 건물에 대한 부가정보나 모델링을 사용하지 않고 DTM과 카메라 외부표정요소만을 이용하여 폐색영역을 탐지한다는 점에서 이러한 연구들과 차별성을 가진다.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.18 no.1
• /
• pp.51-57
• /
• 2000

With the emergence of laser mapping systems, higher resolution DTM of urban area can be acquired and can be used to generate precise orthoimage. But, when the conventional orthoimage generation methods are applied to the area containing features with height difference such as cliffs, bridges, banks. elevated highways and buildings, they cause problems such as occlusion and double mapping. Therefore, this study proposes a new algorithm by modifying and refining conventional orthoimage generation methods. With this algorithm, areas which have occlusion are detected from the base image using camera orientation parameters and DTM. Also, detected areas are restored using alternative images which does not have occlusion in that area. This study can be distinguished from the other studies in the aspects that the proposed algorithm in this paper doesn't need information on building and that uses DTM data and orientation parameters.

• Proceedings of the KSRS Conference
• /
• 2009.03a
• /
• pp.13-17
• /
• 2009

고해상도 위성영상의 분류 기술은 최근 가장 활발히 연구되고 있는 분야 중 하나로 텍스쳐(texture), NDVI, PCA 영상 등 다양한 전처리 정보들을 추출하고 이를 멀티스펙트럴 밴드와 조합하여 분류 정확도를 높이는 기술을 개발하는 연구들이 주를 이루고 있다. 고해상도 위성영상에서 건물의 그림자와 옆벽면의 폐색 지역은 개체 추출 및 분류를 방해하는 주된 요인이 되며, 다양한 형태와 분광특성을 갖는 개개의 건물은 자동 분류 과정을 통해 제대로 식별되지 않는다는 한계를 갖는다. 이에 본 연구에서는 KOMPSAT-2 단영상으로부터 효율적으로 건물 정보 및 토지피복을 분류하기 위하여, 추출된 건물 정보를 바탕으로 건물의 그림자와 폐색지역을 보정한 후 비건물 지역에 대한 분류를 수행하여 분류 정확도를 높이고자 하였다. 우선 삼각벡터구조 기반의 반자동 인터페이스를 이용하여 건물의 3차원 모델 및 그림자 영역을 추출하고 이로부터 추출된 그림자 영역을 효과적으로 보정하기 위해 반복 선형회귀 연산을 이용한 그림자 보정을 수행한 후 inpainting 기법을 건물 폐색영역 복원에 적용하여 영상의 품질을 향상시켰다. 이러한 과정을 통해 도심 지역의 영상 분석에 있어 가장 큰 오차를 일으키는 인공물의 그림자와 폐색에 의한 오차를 최소화한 후 분류에 적용하여 이를 보정 전 영상을 이용한 분류 결과와 비교하였다.

• Proceedings of the KSRS Conference
• /
• 2008.03a
• /
• pp.59-62
• /
• 2008

고해상도 위성영상의 관측시, 위성센서는 보통 지표면으로부터 어느정도의 기울기를 갖는 상태에서 촬영이 되기 때문에 영상 내에서 건물은 지표면에 누워있는 형태로 나타나게 된다. 때문에 건물의 옆벽면 및 지붕에 의해 지표의 일부가 가려지게 되는데 이를 건물에 의한 폐색영역이라 한다. 이러한 폐색영역은 건물의 기복오차가 제거된 정사영상에서는 검게 비어있는 상태로 남게 되며 시각적으로나 영상판독시 불편을 초래하여 위성영상을 베이스 맵으로 사용하기 어렵게 하는 요인이 된다. 이러한 폐색영역을 보정하기 위해서는 일반적으로 동일 영역에 대한 두 장 이상의 영상을 이용하여 폐색 지역을 채워넣는 작업을 수행하나, 이 방법은 위성영상 구입 및 처리 비용에 대한 부담이 커 실제로 자주 사용되지 못 한다. 본 연구에서는 고해상도 위성 단영상의 건물에 의한 폐색영역에, 주변 화소값들의 분광 및 기하학적 특성을 이용하여 복원하는 기술인 inpainting 기법을 적용하여 그 보정 결과를 평가하고 활용 가능성을 검증해보고자 한다.

• Proceedings of the Korean Society of Surveying, Geodesy, Photogrammetry, and Cartography Conference
• /
• 2003.10a
• /
• pp.237-242
• /
• 2003

During the past, digital orthophoto is generated for rural area or low resolution image, because the accurate extraction of DEM is difficult for urban area. But, nowadays, high resolution DEM by ALS system starts to become available for urban area, so the importance of large scale digital orthophoto generation becomes increasing. In this paper, we propose and describe effective algorithm for detecting occlusion area and not only restoring occlusion area but also processing null pixels by occlusion area for minimizing the heterogeneity of digital orthophoto. With proposed algorithm, we detected occlusion area due to height of structures such as buildings, bridges, etc, and restored occlusion area using reference image. Also, The homogeneity of generated digital orthophoto was improved by using brightness correction.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.40 no.3
• /
• pp.177-185
• /
• 2022

A method of restoring the occluded area was proposed by referring to images taken with the same types of sensors on high-resolution optical satellite images through deep learning. For the natural continuity of the simulated image with the occlusion region and the surrounding image while maintaining the pixel distribution of the original image as much as possible in the patch segmentation image, CycleGAN (Cycle Generative Adversarial Network) method with ConvNeXt block applied was used to analyze three experimental regions. In addition, We compared the experimental results of a training patch size of 512*512 pixels and a 1024*1024 pixel size that was doubled. As a result of experimenting with three regions with different characteristics,the ConvNeXt CycleGAN methodology showed an improved R² value compared to the existing CycleGAN-applied image and histogram matching image. For the experiment by patch size used for training, an R² value of about 0.98 was generated for a patch of 1024*1024 pixels. Furthermore, As a result of comparing the pixel distribution for each image band, the simulation result trained with a large patch size showed a more similar histogram distribution to the original image. Therefore, by using ConvNeXt CycleGAN, which is more advanced than the image applied with the existing CycleGAN method and the histogram-matching image, it is possible to derive simulation results similar to the original image and perform a successful simulation.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.38 no.6
• /
• pp.655-661
• /
• 2020

SGM (Semi Global Matching) can be used to find all the conjugate points between stereo images. Therefore, it enables high-density DSM (Digital Surface Model) production from high-resolution satellite images. However, water, shadows, and occlusion areas cause mismatching of the surrounding points in this method. Particularly, in buildings with large-parallax and elongated-shapes such as a Korean style apartment, it is difficult to reconstruct the 3D building even if the SGM method is applied to a high-resolution 50cm satellite image. This study proposed and performed the SGM technique with a classified image and an edge image from the IKONOS-2 satellite stereo-image with a 1m resolution to produce DSM. It was compared with the DSMs from the general SGM and the high-density ABM (Area Based Matching) matching of ERDAS software. The results of the apartment DSM by the proposed method were the best in the test area. As a result, despite the image having a resolution of 1m, the outline of the building DSM could be expressed more clearly than the existing method.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.38 no.4
• /
• pp.363-373
• /
• 2020

During last decades numerous studies generating orthoimage have been carried out. Traditional methods require exterior orientation parameters of aerial images and precise 3D object modeling data and DTM (Digital Terrain Model) to detect and recover occlusion areas. Furthermore, it is challenging task to automate the complicated process. In this paper, we proposed a new concept of true orthoimage generation using DL (Deep Learning). DL is rapidly used in wide range of fields. In particular, GAN (Generative Adversarial Network) is one of the DL models for various tasks in imaging processing and computer vision. The generator tries to produce results similar to the real images, while discriminator judges fake and real images until the results are satisfied. Such mutually adversarial mechanism improves quality of the results. Experiments were performed using GAN-based Pix2Pix model by utilizing IR (Infrared) orthoimages, intensity from LiDAR data provided by the German Society for Photogrammetry, Remote Sensing and Geoinformation (DGPF) through the ISPRS (International Society for Photogrammetry and Remote Sensing). Two approaches were implemented: (1) One-step training with intensity data and high resolution orthoimages, (2) Recursive training with intensity data and color-coded low resolution intensity images for progressive enhancement of the results. Two methods provided similar quality based on FID (Fréchet Inception Distance) measures. However, if quality of the input data is close to the target image, better results could be obtained by increasing epoch. This paper is an early experimental study for feasibility of DL-based true orthoimage generation and further improvement would be necessary.