Semyung university, Dept. of civil engineering
Background and Objective
Practical use of 3D Perspective Image Map in Construction field is on an increasing trend after going through many changes in creation method and technology and ever since the composition of various Geographic Information became possible by Ortho-Projection technique using Digital Map and Satellite Images.
Since the construction planning and designing through the composition of pictures, image data and three-dimensional presentation of Digital Topographic Map is possibly carried out, the application of Spatial information system to the construction field is being considered as a great solution for a new construction planning technique. Because High resolution Remote Sensing Image and GPS Information to substitute for the existing Digital Topographic Map and Thematic Map have been already provided and it seems that field application to various fields is highly possible, in this study, we collected detailed topographic information about the routes of expressway and local road which is now construction-expected and carried out Stereo Terrain Analysis of the construction-expected planned-routes to supply basic topographic spatial information for the most reasonable optimum route selection.
To get local environmental information about the chosen districts, you can collect topographic map, air-photo and satellite image data and create 3D Perspective Image Map using that information and then make use of the 3D Perspective Image Map for analysis of local environment. To explain the steps that we took in this study briefly; we firstly created Ortho-Image of Yangsan and yongwol using satellite images from Arirang 1, LANDSAT TM(America), SPOT(france), IKONOS(USA) and topographic map made by National geographic institute (NGI) and other information about the target district. And created 3D Perspective Image Map using the Ortho-Image and DEM. And then performed simulation using the multidirectional 3D Perspective Image Map.
After all, the objective of this study is to find the optimum route and carry out the most reasonable construction planning/designing through this whole procedure.
3 Dimensional Image Creation
For the 3D Terrain Analysis of satellite images, collect satellite images of target district according to the purpose and property of objects and use them after color-composite. Differentiate Road network files from Contour Line files of Digital Topographic Map which was made to get DEM(Digital Elevation Model) of target district and use the suitable software for processing of collected topographic information. The three requisites which should be primarily prepared for this study are as follows:
- the initial file of the HDF formated satellite image of target district(Yangsan) captured by EOC sensor of Arirang 1, and the IMG formated satellite image of the same district captured by TM sensor of LANDSAT 6, IKONOS for display and stereo image generation at yongwol railway.
- The files of 1:5,000 Topographic Map made by National geographic institute (NGI) as a digital map and road’s planned route data;
- CAD files of road’s planned route of the target district.
Then survey the suitable s/w solution for integration of Vector and Raster. In this study, PCI Geomatics is adopted as the digital image processing s/w for Spatial analysis
After preparation of images and data which is suitable for the purpose of this study, we’ve got to find the precise geographic coordinate of target district on the images. For this reason, we carried out Geometric Correction using GCP(Ground control point) to make three images, which are different from each other in resolution, fit into Geographic Coordinate and after that, carried out Close Ortho Correction using TM Coordinate, a geographic coordinate especially used for the current construction design. (1) Projection on Geographic Coordinate
The basic map projection method for Ortho Correction Image Mapping is TM E002 Projection. And that is the map projection method being used by National geographic institute (NGI) for mapping.
(2) Data Input
The primary satellite image used in this study is the image from Arirang1 (hereinafter KOMPSAT). Purchase HDF formatted Pancro Band(1 channel, black-and- white) captured by EOC sensor and calculate the orbit information of the moving satellite.
(3) DEM Creation
Convert DXF formatted file of 1/5,000 Topographic Map to PIX formatted file. It’s because the format of Ortho Engine, the image processing s/w, is PIX. After format conversion, you can create DEM.
Picture 1. The results of 10m interval DEM creation by using vector of Contour Lines of target district (Yangsan, Kyongnam)
Picture 2. The results of 5m interval DEM creation by using vector of Contour Lines of target district(Yongwol)
Figure 1. DEM of RGB expression
(4) Ortho Correction Image Mapping)
1) Collecting GCP(Ground Control Point)
To get precise ortho correction image, we collected coordinate values of GCP corresponding to specific points of satellite image using files of 1/5,000 Topographic Map and DEM files. In this study, once we found the location of a fixed construction such as a building roof, end point of a bridge and a corner of road, we could get TM coordinate and the altitude of the construction on digital map. Table 1. The results of Collecting GCP
Table 1. Matching point of image and map for GCP
|GCP ID||Column(x)||Line(y)||X (m)||Y (m)||(m)|
To confirm the matching accuracy of these GCPs, we evaluated the accuracy of GCP which was used for final close correction by using RMSE(Root mean square error) method. As the results, RMSE value of the above table turns out to be 5.08 m(0.74 pixel). In case the RMSE value for final one pixel is about 8 m, this result can be considered to be within the allowance. Since this RMSE value is below than the Spatial Resolution of Arirang satellite image before color-composite(6.6 m/1pixel), it’s definitely considered to be within the allownace
2) Correction Estimation
To create Ortho-Projection Image, we carried out the following two corrections:
- BAM (Bandle Adjustment Method): Estimation of exterior orientation using GCP;
- Algorithm of SOM(Satellite Orbit Modeling): Correction of systemic distortion.
Picture 2. The result of overlapping Road network on top of the color-composite image after Ortho-Correction.
Figure 2. Road network overlaid on correction image
3D Perspective Image Mapping Using Satellite Image
1. Preliminaries to Perspective Image Mapping
For Perspective Image Mapping, we used the image creation techniques based on Projection View Method. As preliminaries to Perspective Image Mapping, we prepared information like table 2. and set up fusion-image mode to express textures of terrain softly.
TABLE 2. Image data of perspective map
|RGB Image||Color composite Channel of Kompsat and Landsat Images|
|DEM Image||DEM Image Channel Generation from Digital Contour map|
|Vector Map layers||Digital Road for Planning Routes|
FIGURE 3. RGB image + vector road
FIGURE 4. DEM + vector road
2. Perspective Image Mapping by Projection View Method
In this study, Perspective Image Mapping model was made through the composition of various files to express in Perspective Map using PCI Modeler. To describe it briefly, the process is as follows; Inputting RGB channel, DEM channel and Vector layer through IMPORT module Creating projected image through Image Projection Algorithm of PSGIMAG module Showing the results on the computer screen through VIEWRGB module Creating the files of 3D Perspective Image Map through EXPORT module.
FIGURE 5. Perspective formation diagram
TABLE 3. Perspective production model
|psgimag-east-45.mod||EAST||From Right Center to Left End Direction|
|psgimag-west-45.mod||WEST||From Left Center to Right End Direction|
|psgimag-south-45.mod||SOUTH||From Bottom Center to Above End Direction|
|psgimag-north-45.mod||NORTH||From Above Center to Left End Direction|
Perspective Image Map, which is supposed to be made through this study, is a 3D Image viewed from 45 angle in the four cardinal directions and the detailed features are like table 4.
TABLE 4. Contents by model
|EAST||From Right Center to Left End Direction||6000m||45°||60°||Real|
|WEST||From Left Center to Right End Direction||6000m||45°||60°||Real|
|SOUTH||From Bottom Center to Above End Direction||1800m||45°||60°||Real|
|NORTH||From Above Center to Left End Direction||1800m||45°||60°||Real|
(Full-Scale: 1:1 expression without any exaggeration)
3. Result of Perspective Image Map Creation
FIGURE 6. Perspective Image from the east
FIGURE 7. Perspective Image from the west
FIGURE 8 . Perspective Image from the south
FIGURE 9 . Perspective Image from the north
Results and Prospect
This study has been accomplished as a experimental study for 3D Perspective Image Map analysis of terrain along the 16km long section between Yangsan and east Busan which is under consideration for a new expressway construction. Through this study we could have some satisfactory results to carry out multidimensional terrain analysis of remote place using the field survey data of Satellite image, Topographic map and road’s planned route. To carry out Stereo Terrain Analysis using various spatial information of topography, we made Perspective Image Map of districts around construction-expected road from various directions and altitudes and produced moving images of the construction-expected road for virtual driving along the planned route. Therefore, on the construction-planning phase, 3D analysis based on the results of this study can be simulated with various geographical features and facility conditions. 3D Perspective Image Map to carry out stereo terrain spatial analysis in the four cardinal directions can be considered as a great solution for a new construction planning technique and for the most reasonable optimum route selection.
Lately, as various sensors of satellite and aircraft made collecting of spatial information easier and we can easily get spatial information with high resolution, and more than 3D spatial analysis techniques and time-spatial analysis techniques are being developed, we can expect to make good use of much more realistic Perspective Image Map in real life.
- Yeon, Sangho and. 2001. Introduction to Remote Sensing, Goomi Book Center Publisihng.
- Yeon and Lee, 1994. An Introduction to GIS and Exercise, Hanwol Academy Publishing.
- American Society of Photogrammetry. Manual of Remote Sensing. 2nd Edition
- ASPRS. 1989. Non-topographic Photogrammetry 2nd Edition
- C. H. CHEN. Information Processing for Remot Sensing. World Scientific.
- Lillesand, Kierer. Remote Sensing And Image Interpretation. WILEY
- Martin David Adams. Sensor Modelling, Design and Data Processing for Autonomous Navigation. World Scientific
- PCI Geomatics. 2001. Geomatica Software manual.
- Robert H. Arnold. Interpretation of Airphotos and Remotely Sensed Imagery. Prentice Hall.