Poster Session 1
Engineer, Water Conservancy agency No.501, Li-Ming rd Taichung Taiwan
After the serious 921-Tsau-Lin Earthquake, it's necessary and urgent to know the landslide condition and to handle the effectiveness of the silted earth mass on its upper and lower stream channel. Those data have been urgently concerned about engineering field. To integrate the aerial photogrammetric photograph, GIS and image processing software, and modeling 3-D image, drown range delineation, the overtopping height calculation and the storage water volume etc., we can finish the preprocessing operation. Next, to create the 10m*10m resolution DTM of the site before and after the earthquake. Finally, to overlap that coverage's we have finished, we can get the reliable volume of landslide.The purpose of this article is to make a complete introduce the results and some important procedures how we utilized the scientific technique of aerial photogrammetry, image processing and geographic information system etc. Finally, we made a real and complete record about the important digital information of Tsau-Lin landslide.
A largest earthquake of this century in Taiwan occurred on September 21, 1999 and caused deaths of more than 2,000 human lives and loss of billions dollars worth in houses, buildings, roads, bridges and dams. This disastrous earthquake also induced 2,365 landslides and soil mass movements with total area of 14,000 ha. After the serious earthquake of central Taiwan, the biggest one, landslide dammed the Chin-Shui-Chi River and formed Tsau-Lin reservoir, it's necessary and urgent to know the landslide condition and to handle the effectiveness of the silted earth mass on its upper and lower stream channel and the mud disaster that may caused owing to the silted earth mass. Those data have been urgently concerned about by both research institute and engineering field such as: the location, area and the volume of the silted earth mass, the drown range of water, the over flow height, and the storage water volume. To integrate the aerial photogrammetric photograph, GIS and image processing software, and modeling 3-D image, drown range delineation, the over flow height calculation and the storage water volume etc., we can finish the preprocessing operation. Next, to create the 10m*10m resolution DTM of the site before and after the earthquake. Finally, to overlap that coverage's we have finished, we can get the reliable volume of silted earth mass.
The purpose of this article is to make a complete introduce the results and some important procedure how we utilized the scientific technique of aerial photogrammetry, image processing and geographic information system etc. and quickly to handle the field situation and processed the relative digital data. Finally, we made a real and complete record about the important digital information of Tsau-Lin landslide. The emergency treatment project to reduce the disaster can be determined helpfully and rapidly.
2.The history of Tsao-Ling landslide.
According to the documents, the Tsao-Ling landslides were decided by seismic and intense rainfall factors, the another factors were topography, soil, geology, and active fault zones.
An unusual series of events illustrating the formation and destruction of a landslide dam has been documented for the Tsao-Ling landslide in central Taiwan. This case demonstrates just how complex natural dam processes can be. In June 6,1862, a major earthquake triggered a landslide that dammed the Chin-Shui-Chi River. In 1898, the natural dam failed for unknown reasons. In December 17, 1941, a major earthquake formed another landslide dam, 140 m high, at the same location. In August 10, 1942, heavy rainfall caused reactivation of the landslide and the natural dam increased in height form 140 to 217 m. In May 1951, several days of intense rainfall led to the overtopping and failure of the natural dam. In the subsequent flood, 154 people were killed and 564 homes damaged. On August 15,1979, heavy rainfall again activated the landslide, which dammed the river with a natural barrier 90 m high. Heavy precipitation continued, and 9 days later, the landslide dam was overtopped and failed, causing severe flooding (Costa and Schuster 1988).
In September 21,1999, Richter scale 7.3 earthquake formed landslide dam again, the detailed landslide data were gathered rapidly by integrating aerial photogrammetry and GIS. The main data of Tsao-Ling Landslide are shown follow :
a. Deposit Height of Upstream:50m (EL.539.6m),
b. Area of Upstream Watershed:162Km2 ,
c. Landslide Area:620ha,
d. Deposit Distance:5Km,
e. Landslide Volume:126 million M3(Cut) and 150 million M3(Fill),
f. Lake Volume:46 million M3 ,
g. Overtopping Elevation:EL.539.6m,
h. Second pond volume:4.99 million M3 for EL.520.4 overtopping, 3.4 million M3 for EL.515 overtopping.
3. Processing of created DTM of before and after landslide
In order to compare the surface deformation before and after landslide, it's rapid method to create the high resolution Images and DTM of the site, before and after the earthquake. We can compare, overlap, and calculate with these digital data conveniently and precisely. The detailed flow chat is shown as figure 1.
Poster Session 1
Photogrammetry was invented in 1851 by Laussedat, and has continued to develop over the last 140 years. The development of photogrammetry has passed through the phases of plane table photogrammetry, analog photogrammtry, analytical photogrammetry, and has now entered the phase of digital photogrammetry (Konecny, 1994). In analytical photogrammetry, the computer replaces some expensive optical and mechanical components. The resulting devices were analog/digital hybrids. Analytical aero-triangulation, analytical plotters, and ortho-photo projectors were the main developments during this phase. Outputs of analytical photogrammetry can be topographic maps, but can also be digital products, such as digital maps and DEMs. Digital photogrammetry is photogrammetry as applied to digital images that are stored and processed on a computer. Digital images can be scanned from photographs or can be directly captured by digital cameras. Many photogrammetric tasks can be highly automated in digital photogrammetry ( e.g., automatic DEM extraction and digital ortho-photo generation) (IMAGINE Tour Guide 1998). A strip of photographs consists of images captured along a flight line, normally with an overlap of 60%~70%, with a sidelap of 20%~30%(figure 2, figure,3). Especially for mountain area the overlap and sidelap more will be better. Strips and exposure points have been calculated so that the specifications concerning overlapping and total ground coverage is being fulfilled.
Figure 1: processing flow chat
Figure 2:strip of photographs desigh
Figure 3?overlap and sidelap area (IMAGINE Tour Guide 1998)
In July 1975, the Executive Yuan approved a project for producing large -scale photo base map covering all of Taiwan. Aerial photo and mapping works are executed by the Agricultural and Forestry Aerial Survey Institute , Taiwan Forestry Bureau( RU, 2000). So, we have many existing base maps and aerial photos about the Tsao-Ling landslides to be applied. When the earthquake occurred, the Agricultural and Forestry Aerial Survey Institute have captured full disaster area color photo in late two or three days by good weather. The Tow-Lin survey company have completed creating DTM and Maps in two days rapidly. Next we used these digital data integrating GIS and RS tools to obtain landslide area, volume, water volume, 3D view, and overtopping elevation, etc. Image measurement and calculations must be done rapidly and effectively in the emergency situation. Therefore there are needs to automate the measurements and to develop an effective user interface.
Each photograph or image that is exposed has a corresponding image scale associated with it. The image scale expresses the average ratio between a distance in the image and the same distance on the ground. It is computed as focal length divided by the flying height above the mean ground elevation.
Sh = f/H-h Sh:Photo-Scale of Point h
f:Focus length of Camera
This emergency tasks of aerial photographing with zeiss RC10 cameras, focus length 135.04mm, 1/10000 photo scale. The cameras that are used for aerial photographing are designed exclusively for that purpose. According to the single hardcopy color films that is approaching the real sight, we can interpreter and check the magnitude and area of landslide site rapidly, and we can also take a simple measurement for obtaining the location, situation, and disaster area information with the photo scale instantly.
3.2 Images Scan: mapping and ortho-photo
Though digital cameras have developed much the latest years they still cannot compete with the resolution of analogue cameras when it comes to aerial images. Thus the images has to be scanned to digital format before they can be used in digital photogrammetry applications. Photogrammetric quality scanners are special devices capable of high image quality and excellent positional accuracy. The required pixel resolution varies depending on the application. Aerial triangulation and feature collection applications often scan in the 10 to 30 micron range. Orthophoto applications often use 15 to 30 micron pixels. Color film is less sharp than panchromatic, therefore color ortho applications often use 20 to 40 micron pixels (IMAGINE Tour Guide 1998). Scanning aerial color film with high resolution following the procedure of mapping images, ortho-photo, and triangulation, stereo-pair model calculation by Tow-Lin company with high technology and computer, it's very efficiently. These task facilities as follows:
- VEXCEL Ultra Scan 5000?20 micron pixels resolution
- Leica SD2000 analytical
- VirtuoZo Digital Images Workstation
- SGI Octane work station with 144G Disk Array
3.3 Results of Arial Triangulation
Owing to Earthquake-caused surface ruptured, and we can not survey the exact coordinate by GPS at the landslide site immediately, so we used 1998 (1/5000) 1980 (1/10000) photo base map to calculate the control points of the coordinate and height for triangulation. The results of Arial Triangulation are shown at Table 1.
Table1:Results of Arial Triangulation
TM2 error Number of
Points X-axis Y-axis elevation X-axis T-axis elevation A1 213357.202 2609402.837 323.592 -2.798 2.837 0.142 A2 212194.921 2609081.358 779.789 0.921 1.358 -0.311 A3 211974.992 2609272.138 756.735 2.992 -1.862 -0.615 A4 212692.065 2607412.883 427.478 0.065 0.883 -0.122 A6 217875.808 2609152.702 764.273 A7 218260.828 2607654.870 522.190 -1.172 -5.130 -0.160 A8 217933.711 2608797.932 715.870 5.711 -2.068 0.770 A9 217967.963 2609634.568 856.824 -0.037 2.568 0.324 A10 216852.131 2607677.368 570.688 0.131 -2.632 0.588 A14 213390.126 2606605.351 656.569 0.126 -2.649 0.569 A16 212066.016 2609884.016 916.938 A17 213910.384 2610729.050 330.253 -1.616 1.050 0.253 A18 218056.550 2606098.051 710.411 A19 216305.210 2606227.056 1251.897 1.210 3.056 -1.103 A20 218156.072 2610030.099 960.040 -3.928 2.099 0.040 A21 220518.394 2607676.490 573.625 -1.606 0.490 -0.375
3.4 Created DTMs of before and after landslide
With the 10- m pixel size resolution that is enough to demonstrate the deformation of the landslide surface which is better than the existing DTM of 40- m resolution of Taiwan. For improving the accuracy and efficient, we have created the DTM of before and after landslide with 1998, 1/5000 photo base maps, with the same coordinate system and pixel resolution.
Poster Session 1
We used these digital data integrating GIS and RS tools to obtain landslide area, volume, water volume, 3D view, and overtopping elevation, etc. Image measurement and calculations must be done rapidly and effectively in the emergency situation, by using IMAGINE VITUAL GIS and ARCVIEW GIS?3D model. The results of cut and fill area of the landslide shown as Figure 4. Overtopping Elevation and Water Volume can be calculated with RS tools, results shown as table 2, table 3 and figure 5. Applied the 3-D view and fly through also.
Figure4 calculation of the volume of landslide by Arcview 3D Analysis
( Blue : cut, Red :fill )
Table2: Modeling prediction of Volume and overtopping elevation of Dammed lake
Water elevation Volume(m3) 527 28430.0 528 29706.7 529 31020.6 530 32360.6 531 33726.3 532 35123.8 533 36573.1 534 38083.4 535 39603.1 536 41161.8 537 42741.9 537.5 43543.6 538 44352.8 538.5 45169.2 539 45992.8 539.6 46824.3 overtopping
Table3: Modeling prediction of Volume and overtopping elevation of second pond
Water Elevation Volume(m3) 500.00 454,889 505.00 1,123,360 506.00 1,300,289 507.00 1,489,370 508.00 1,693,369 509.00 1,908,209 510.00 2,132,709 511.00 2,368,199 512.00 2,614,969 513.00 2,872,049 514.00 3,140,039 515.00 3,419,429 Overtopping after Channeled 516.00 3,708,899 517.00 4,011,479 518.00 4,325,979 519.00 4,652,739 520.00 4,993,049 520.40 overtopping Original
Figure 5 modeling submerged area and water volume by Erdas Imagine Virtual GIS water layer model
Figure 6 landslide full 3-D view
Poster Session 1
5.1 Because of the complex and variation of landcover in the mountainous area, differences between digital surface model (DSM) and digital terrain model must be suitful treated (Chen, 1999). Owing to the surface and feature of landslide area are clear, bright, and outshine, it's helpful to automatic feature based matching. Especially, the elevation will be accuracy more than other tree or creek area so that the widely Tsao-ling landslide area elevation measurement is very accuracy for calculating the overtopping height.
Photogrammetry is suitable and available for large disaster area mapping. In this case it just takes two days to complete entire landslide area mapping tasks.
The aerial photos were digitized with high resolution, and automatic aerial triangulation, automatic matching, 3-D stereo mapping were performed to apply digital photogrammetric mapping techniques to improve the accuracy of DTM while maintaining high-quality product and low-cost.
The results of photo automatic matching depend on photos likelihood, less ratio of base line and flying height, tone, and scanned resolution (Chen, 1999). The relations between Photo-Scale and Map-Accuracy are shown as Table 4.
Table 4: Relations between Photo-Scale and Map-Accuracy ( Petrie 1990)
Photo scale resolution
@ 40 1 p/mm Altitude of
flight(m) Map scale Photo/map Interval of
Contours(m) 1:3000 0.075 m 450 1:500 6x 0.5 1:5000 0.125 m 450 1:1000 5x 1 1:10000 0.25 m 1500 1:2500 4x 2 1:25000 0.625 m 3750 1:10000 2.5x 5 1:50000 1.25 m 7500 1:50000 1x 10 1:80000 2.0 m 12000 1:100000 0.8x 20
5.4 The Taiwan Datum 1997 (TWD97) coordinate system is not used in this plan at the pre-period processing. The coordinates of control points are surveyed by GPS after 921 Chi-Chi earthquake. But in the late processing of the Tsao-lin digital data, we have added GPS data and have improved the accuracy.
6.1 We have successfully and rapidly integrated RS?Aerial photogrammetry?GIS?GPS to obtain digital data about the landslide. By Using produced DEM data, and arcview 3D Analyst model and Erdas Imagine Vitural GIS. We have estimated the total landslide area about 620 ha, a volume of about126 million m3 (figure 1), storage water volume of the upstream about 46 million m3 with overtopping EL.539.6m, the second pond volume about 3.14 million m3 with overtopping EL.515m.
6.2 In addition to orthoimages, photogrammetry can also provide other geographic information such as a DTM, topographic features, and line maps reliably and efficiently. Photogrammetry produces accurate and precise geographic information from a wide range of photographs. Any measurement taken on a photogrammetrically processed photograph reflects a measurement taken on the ground. Rather than constantly go to the field to measure distances, areas, angles, and point positions, photogrammetric tools allow for the accurate collection of information from imagery. Photogrammetric approaches for collecting geographic information save time and money, and maintain the highest accuracy.
- Taiwan provincial water bureau (1979) The history of Tsao-Ling landslide.
- Ho, wei-sin (1995) Aerial photogrammetry.
- Yeong-Kuan Chen, Jihn-Fa Jan, Yin-Lin Wu, Kuen-Sheng Yeh (1999) Automatic Matching of DTM for Forested Mountain Areas The 18th Symposium on survey and applications.
- IMAGINE Tour Guide (1998) Photogrammetry and IMAGINE Orthobase, Chapter 2, pp.7-50.
- Petrie, G..(1990), Photogrammetric Methods of Data Acquisition for Terrain Modeling, in Terrain Modeling in Surveying and Civil Engineering, Eds. G. Petrie, T. J. M. Kennie, Whittles Publishing, pp.26-48.
- Ru, j. c. (2000), Production of photo base map in Taiwan, The 4th Symposium on GPS Technology, pp.181-193
- Costa and Schuster (1988), The formation and failure of nature dams, Geological Society of America Bulletin, v.100, pp 1054-1068.