International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
  
  
2.3 LIDAR Data 
In Tokyo, the LIDAR data acquired are stored in the form of 
“library” by the private enterprise, and sold to the general 
public. In this study, we used the library data of different 
acquisition density obtained by rotary wing and fixed wing. 
The data of rotary wing are acquired by ALMAPS owned by 
Aero Asahi Corp. on July 26 to 27, 2002 which have been 
processed as DSM (Digital Surface Model) points group and 
DEM (Digital Elevation Model) points group of 1 m mesh. As 
for the details of measurement, the flying altitude of airplane 
was 1,000 m, pulse rate was 25 kHz, scan rate was 25 Hz and 
foot print was 0.013 m. The measuring accuracy of scanner is 
+/- 50 em for horizontal direction and +/- 15 cm for vertical 
direction (Table 2). 
The data of fixed wing are acquired by RAMS owned by 
Kokusai Kogyo Co., Ltd. on September 24 to 25, 2001 which 
have been processed as DSM points group and DEM points 
group of 2 m mesh. As for the details of measurement, the 
flying altitude of airplane was 2,590 m, pulse rate was 24 kHz, 
scan rate was 24 Hz and footprint was 0.66 m. The measuring 
accuracy of scanner is +/- 30 cm for horizontal direction and +/- 
15 em for vertical direction (Table 2). 
In this paper, LIDAR data of 1 m mesh is called the “high- 
resolution data” and LIDAR data of 2 m mesh is called the 
“low-resolution data.” 
3. METHODS 
3.1 Extraction of Area Covered by Vegetation 
The flow of this study is shown in Figure 2. First of all, we 
extracted the areas covered by vegetation by calculation of 
NDVI (Normalized Differential Vegetation Index) using the red 
band and near infrared band of IKONOS. The threshold value 
between the areas covered by vegetation and not covered by 
vegetation was specified as NDVI = 0.1. 
3.2 Extraction of Height of Structures on the Ground 
We made interpolation by TIN (Triangular Irregular Network) 
to the DSM points group and DEM points group of LIDAR data 
and converted them to grid data of 1 m mesh and 2 m mesh. In 
addition, we added subtraction processing to DSM grid and 
DEM grid. The processed result is hereinafter called the DHM 
(Digital Height Model). DHM is the model which indicates the 
height of ground structures and trees. 
3.3 Extraction of Tree Height 
We separated the trees and ground structures by clipping DHM 
in vegetation cover area extracted by IKONOS. This processed 
result is hereinafter called DCHM (Digital Canopy Height 
Model). DCHM is the model which indicates the height of tree 
crown. DCHM of | m mesh is shown in Figure 3, and DCHM 
of 2 m mesh in Figure 4. 
3.4 Verification of Measuring Result of Tree Height 
We verified the measuring accuracy of tree height by 
comparing the actually measured and DCHM for each study 
514 
sites. In addition, we discussed the relationship among the 
measuring error, tree types and resolution. 
  
— Meters 
600 
200 400 
Figure 1. Study sites of this study 
  
  
  
  
  
  
  
  
  
  
  
  
St.No.] N aveH (sdH) | aveD (sdD) | N.S TT 
St.1 45 10.8(5.0)] | 35.0 (17.2) 8| b,m,d 
St.2 18 16.5 (3.2)] 49.4 (12.5) 3 b, u, s 
St.3 21 20.6 (5.3)] 55.1 (15.4) 3 c, u, d 
St.4 14 13.8.(4.2)1 1.5 (19.1) 2 b,u,s 
St.5 27 13.5(3.7)]  56.1(172) | b, u, d 
N: number of tree, aveH: average height (m), sdH: standard 
deviation (m), aveD: average diameter at breast height (cm), sdD: 
standard deviation (cm), N.S: number of tree species, T.T: tree 
  
  
  
  
  
  
types (b=broad-leaf, c=coniferous, m=multi-storied forest, 
u=uniform forest, d=dense, s=sparse) 
Table 1. Detail of ground truth result 
Type Name Level S.R Date 
Satellite Image | IKONOS | Multispectral 4m 2000/9/21 
DSM Im |2002/7/26 -27 
AL oS 
LIDAR LARS DEM Im 1|2002/7/26 - 27 
í RAMS DSM 2m 2001/9/24 - 25 
DEM 2m __12001/9/24 - 25 
  
  
  
  
  
S.R: Spatial resolution 
Table 2. Data used for this study 
  
/ IKONOS DSM(Im) / DSM(2m) 
ries / 
DEM(Im) DEM(2m) 
Y : v Y 
| Calculate NDVI | | Subtraction Subtraction 
i | Processing Processing 
| 
T v i v ; Y i 
| | Green Cover | | | DHM(Im) | | | DHM(2m) | 
|. | Li kd LI ; 
| 
er res A WI 1 
> Clipping Processing 
| 
[Ground Truth | r3 y : s Y 
|, Rest | | | DCHM(Im) | | DCHM(2m) 
| (Tree Height) I ; 1 
A : I 
: Y Y | 
> Verification | 
| Comparison of Accuracy 
Figure 2. Flowchart of this study