ISPRS, Vol.34, Part 2W2, “Dynamic and Multi-Dimensional GIS”, Bangkok, May 23-25, 2001 
90 
Fig.4 GPS technique to collect spatial information 
Fig.5 GPS technique to monitor the subsidence 
situations. IDCS has interface with each methods and can 
process their raw observing data, and can provide GIS with 
processed results by uniform format, generally the set of 3D 
coordinates (X, Y, Z) of all observed points. Fig 6 is the structure 
of IDCS. 
With the IDCS, we can get the 3D information used for 
simulation and representation of subsiding land. 
4 3D DATA STRUCTURE 
3D data structure is the basis of simulation and representation. 
We have already given some discussions on 3D data structures 
used to mines taking into account of properties and applications 
in mines. Here we would propose some useful data structure for 
representation subsiding land. 
Fig.6 the formation and structure of IDCS 
4.1 Surface Representation Using TIN and DEM 
Subsiding land can be simulated by two means: one is 
expressing the land surface according to elevation of every 
points, and the other is simulating the dynamic process of mining 
subsidence including deformation of land surface and movement 
of rock from real 3D space. For the former, TIN and DEM can be 
used, and integrated data structure for the latter. 
Though TIN and DEM are not a real 3D representation, they are 
used widely to represent the surface by 2.5D because the 
algorithm is simple, data is easy to organize and collect, and use 
is convenient. By (X, Y, Z) of all the points collected by IDCS, 
TIN and DEM can be generated easily by GIS software, and 
they can be transformed with contour and each other. Based on 
DEM, some simple measurement and computation can be done. 
4.2 3D Simulations Using 3D Integrated Data Structure 
For complex simulation of mining subsidence, none of the data 
structure can be used effectively and 3D integrated data 
structure is necessary. According to the features of mining 
subsidence, two integrated data structures were proposed as 
follows. 
4.2.1 Integrated data structure of 3D Raster and DEM 
3D raster or Octree is used to represent the rock, coal seam and 
goaf after mining, and land surface is expressed by DEM or TIN. 
Firstly, the deformation and movement of land surface, rock and 
goaf is computed, and then the 3D cell is marked according to 
the generation rule of Octree. Octree and TIN are overlapped 
according to spatial coordinates of some control points In this 
scheme, the position and movement of rock should be computed 
by spatial model and perhaps it is not precise enough. 
4.2.2 Integrated data structure of 3D vector, Octree and TIN 
The goaf by mining often is expressed by 3D vector data 
structure because it can be viewed as a regular 3D space 
composed by vertex, line, arc and face. The rock over goaf is 
represented with Octree and the movement of each voxel can be 
computed by spatial model, and the subsiding land surface is 
represented by DEM or TIN. These three different data