International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
  
(6) Slight Shift of Retaining Wall (Concrete Block) 
(7) Crack of Retaining Wall (Masonry) 
(8) Swell of Retaining Wall (Masonry) 
(9) Slight Shift of Retaining Wall (Masonry) 
We defined the 9 types of concrete structure problems above as 
*concrete anomaly" in this study. 
5. TEMPERATURE PATTERN OF NORMAL 
MATERIALS AND THE EFFECT BY SHADOW 
In general, the radiant temperature observed by the spaceborne 
/ airborne thermal sensors is easily affected by shadow, 
especially on the high resolution images. Therefore, a shadow 
simulation with DEM could be useful to estimate and remove 
its effect. However, the 50 m resolution DEM available in this 
study is insufficient to simulate the shadow area of the 1.5m 
GSD thermal images. Instead of DEM, 2 TABI images in early 
afternoon and in early morning were used to estimate the 
shadow effect. 
The difference of the radiant temperature (AT) was defined as 
below. 
AT = TEA - Tim (1) 
Where Tg, is radiant temperature from the early afternoon 
(EA) image, Ty is radiant temperature from the early morning 
(EM) image. 
The average AT profile of four pure land cover types (Road, 
Vegetation, Concrete, Soil) are shown in Fig.4. Four shadow 
conditions shown in Table. 3 were sampled for each land cover 
  
14.0 
2f o Ha TUM (ec 45 arte T 
12. 
10.0 HN, 
8.0 
ge 6.0 
E 46 
3 > wom moms me a... += 
2.0 
0.0 mn He te ts am yp Benes ts eu Arm on — 
-2.0 2 1 0 1 2 
-4.0 
Pixel 
(a) Road 
14.0 
120 : 
10.0 
8.0 
© 60 
E 4.0 
"€ Me, m----_- FE -a 
2.0 
0.0 T onem ud Sans auk A E = am mi alm a 
20 2 :i 0 1 2 
-4.0 
Pixel 
(c) Concrete 
——:NS 8M 
type from the images. Each profile was taken from the averages 
of three lines of five pixels long (-2, 2). 
For road, vegetation, and concrete, the AT pattern decreases 
from SM, NS, SB to SA. However, for soil, AT under NS is 
greater than that under SM. That is because the soil in shadow 
was still getting cooler at the data acquisition time for the EM 
image. 
From Fig. 4, it can be seen that the trend of AT for road is 
similar to that for the concrete under all shadow conditions 
except for under NS as same as shown in Fig. 3. The difference 
in AT patterns between road and concrete under NS suggests a 
possibility for the discrimination of the two. AT for soil under 
NS is smaller than the value estimated from Fig. 3. This is 
reasonable as soil retains a certain amount of moisture. 
The descending orders of AT by various conditions are 
Concrete, Road, Soil and Vegetation for NS; Road, Concrete, 
Vegetation and Soil for SM and SB; and Soil, Road, Concrete 
and Vegetation for SA. These orders could be a special 
characteristic of thermal behavior of materials, but the soil 
causes complication and will not be examined in detail here. 
Fig. 5 shows the distribution of the radiant temperature in the 
Table. 3 Shadow Conditions 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
N Shadow 
Code+— : Note 
EA | EM 
NS | N N [No shadow in both observation time 
SM| N Y |Covered with shadow only in EM 
SA M. Y N [Covered with shadow only in EA 
SBIY Y |Covered with shadow in both observation time 
14.0 | 
12.0 
10.0 
rtt WR crews sare Meet ec cc cu eee L] 
8.0 
5 6 
Sw] ee . ae 
> 0 Bis --- -— JA. Br -" 
0.0 
20 2 -l 0 2 
| hee — A++ m am 7A 
40 | 
Pixel 
(b) Vegetation 
14.0 
12.0 
10.0 : 
8.0 
f v-…--..... Wenner meee U Meet n » 
z 6.0 
< 40 
2.0 Be Bm Bi | E 
am a nw — — am am —^ 
0.0 
20 3 - 0 1 2 
-4.0 
Pixel 
(d) Soil 
mn SA === SB 
Fig. 4 Distribution of AT for Four Land Cover Types with Normal Condition 
588 
Intern 
Tem(°C) 
Tem(°C) 
Table. 
Land Cove 
Roa 
Vegeta 
— 
Concr 
IE————— 
Soil 
  
  
four I: 
points 
in Fig. 
temper 
conditi 
widely