Detail Estimation of
with New Ex
ct
Area
om NOAA AVHRR
ta
xel Insi
tion
i
Koganei
Commision VII
TRUCT:
f
a snow
uncI
On
NOAA AVHRR
spati
LANDSAT
logy
from
is
1 INTRODUCTION
Remote sens
snow cove
ctively
en us
r
water
estimation of
LANDSAT
ta are main
of
area
is.
ta
s
tage
to obtain
then it has more
es
tion (IFOV)
is low
t
every 16
LANDSAT
t
winter or
is ra
lope
snow
29SPECIFICATION OF RESEARCH
(1) Data
NOAA AVHRR
ta
LANDSAT TM
ta (Pa
Eo
ta were
se
(2) Test Site
mountains area ar
north-east
t of
Nor
t@:
39
East long : 14
ve
to
tails
g 3
g.4
a
@
Row- 2)
(Fi
26
APR@19
ch is
Ci
Mor
(Fig 3)
@OOsec
00 ec
2)
4
- 141
locat
g 1
g 3
3 HOW TO GET PIXEL INSIDE INFORMATION
jects
g@l
At
rs,
Algorithm of Ana
as snow
Then classified
area were classifi
area.
On the
cover
covered area were re
as 0 % snow covered
o the
, n o snow
pattern exsample of 3
area
91b shows
ent
of
x 3
s window
t
x 3
s re
occupation in 3 x 3
macro
int
pixels area are lar
t pixel inside
average parcentage va
snow cover
as macro one
1
Fig.
the ave
of AVHRR row
or mahal
is generaliz
x 3
Is
Then it is
ssible to
regression model
between fig lc data
fig 1
tao This model is named "MACRO
PIXEL MODEL or MPM"
us
MPM regression
1 is use
for
each pixels to estimate snow covered rate for each pixel
insi
4 PROCEDURE OF ANALYSIS
The proc
es of ana
low
ta
AVHRR
ta
LANDSAT TM
a Geometric co rec ions
purpose computer(FACOM M360)
are operat
wi
of snow cover
area.
b The pr
classi
no
as 100% snow cove
c Snow cove
area are re
snow area are re
as 0 % area
Is
d Account parcentage of snow cove
from primary classifi
e Account average of AVHRR
distance from snow cover
ta and the
f. Sett
up of
ta sets
average
ta
re
ssion ana
is.
and
g. Obtaining the coe
cients of regression
tion.
sett
up
estimation
1 of
xel inside
h. Calculate the snow coverd area estimation map for each
pixels with model formula.
LANDSAT TM data
i. Veri cate
accuracy of estimation wi
of same
teo
4.RESULTS OF ANALYSIS
The re
ts of M.P.M.
estimat
M P.M. estimated res
1 map is
M.P.M. regression is shown in t
Table.l
maps are shown in fig 4.
And
in fig.5.
re
ts of
Results
I T E IVI
IVIP}\i1
Coef.of Corr
SNOW
COV.
Mean Sf1. Err
5.CONCLUSION
a. Cause of errors
Most of errors
are
on
accracy of primary
classification.
The causes of miss-classification were these
two items.
se were topo
cal condition and sun angle
factor, and spect
characteristics of no snow area.
b. Problem for the future
This study regard that primary classified snow area is
100% covered and no snow area is 0 %.
But this postulate is
generous.
Especially
border area does not correspond with
this postulate.
Then it is studied next posutlate
boder
snow area is regarded 50% covered
Border area is estimaed
by mahalanobis generaized distance.
This model does not consider the contents of
no snow
area.
Specially vegitation area has high refrection at IRband in spring season.
It is necessary to consider about these
factors.
1
SNOW
SNOW
SNOW
NO-S! NO-S
r
I
NO-S
I
I
SNOW
NO-S
Fig.lb
SNOW
SNOW
I SNOW
1oo~t
100%
I 0 ~~
100%
0
ii 0-
100%
100% 1100%
0/
/0
Fig.la
AVE. 67fJ6
Snow Covered Area Classified Map
Snow covered area
No snow covered area
The primary snow covered
percentage map
Snow area ~IOO'~
No snow area~O%
Fig.Ie
Average of
snow covered
percentage
::::::.A. v'E. C C T
AVHRR
Make regression
model
Fig.Ie
Fi.g.ld
Average of AVHRR data
or mahalanobis
generalized distance
TM 40
pixels
J
'TlVI
I
Account snow cov
in 40x40 pixel
Fig.If
Estimation map
for each pixels
witb regression
model
100%
60~-6
30?6
70.%
40 Ji'
20,%
100%
90?~
70?6
Fig.lg
TM snow
cov. ?i'ma p
I
t
T'M 40pixeIs
90Ji
70?-6
40~i
8096
40.%
20,9f;
100%
80?~
75?{;
Check the
accuracy
Fig.1
How to get the information of pixel inside
VII ... 418
Scale 1:500,000
Fig.3 The map of test area
19
LEGEND
I
I
-------------------------------
------------------I
I
I
I
I
I
I
I
I
I
I
I
(%)1
percen
cover
I
0
0
20
30
40
50
60
70
80
90
100
1
11
1
31
41
51
61
71
81
91
101
I
I
I
I
I
I
I
I
I
I
I
I
CHARACTER
A S1 NED
#
#
I
I
I
I
1
H
T
I
I
I
*
+
BLANK
BLANK
I
I
I
I
I
I
-----------------------------------~--~-~--~-~-
Snow c
rate map
from NOAA AVHRR
ta
Snow covered rate map
from TM data (
r check)
11111111111111111111111122222222222
77778888888888999999999900000000001
67890123456789012345678901234567890
00000000011111111112222222222333333
12345678901234567890123456789012345
1 ################--H#¥HH+--.* -+IH#T
2 ¥¥#¥##########TT--T#H*T*--++ •• *HHIT
3 ¥HT¥#¥########¥T-+¥##TI++.- •• *¥H#**
4 ITTH¥¥¥########1-1##¥T++*.+.+T#*H.+
5 *-1TH¥#####HT##¥--H##H--¥-+.-T¥H¥1+
6 +**T¥¥######¥###H.T¥#TT+¥- .*¥H1HH+
+TH1###¥###¥####HTH##T++ • • +-T-!¥7
.+TH##TT####HT##¥HT¥¥T.
+- IT
8
.+THH¥H¥###¥T¥###H##TTT.
.+-HT
9
I+HT+
10 • --*THTH¥###H*H¥#####HI+..
..--1* ••
11 -*1¥TH¥##T#¥*1*T#####HT.
.T¥¥¥###¥HH¥T-*T###¥¥H* ••• +11 • • 12
13 .+H#######¥TT*.I¥#HTI¥¥¥- •• +- ++H
-+1
14 HT¥######¥*-- ITT##¥¥###T.
*+
15 ¥########¥¥HT • • 1#######* •
16 ############H. T#######¥-.
17 ##########¥#H*. T#H#####¥+
18 ###########H1. ++#¥#####¥-+
19 ###########H.-1¥¥#T#¥###¥T.
20 #########THH.-1H###HH¥###H.
21 #########HHII+H¥H¥¥¥¥I####¥-.
-+.
22 #########¥+H#TT¥#####*T¥##¥23 ##########¥H#######TH¥+H###T-. .. T*
24 ##########¥#¥H#H###H1H-.¥#HT.. +*
25 ###############1¥#*++*T.I##H+. .. *T
26 ###############H#HH+.IH +H#¥1I
27 ##############H1¥**+ -T. +¥T+*--.
28 #############¥#+¥** -*1 -*HHT+.- • •
29 ##############H +T- .+- -HT**+-- -.+ -H*--+T¥*.
30 ###############1+*
I
.*¥#¥T
31 ############¥###H..
-¥###
32 ############¥¥¥¥1
-*+ +¥###
33 ############11+.-1HHH###
34 ############TH-.
• .1¥¥###
35 ##############¥#¥I.
00000000011111111112222222222333333
12345678901234567890123456789012345
Fig.4
221
222
223
224225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
HH¥¥###########T+*T¥TTT •• +++~+I1TH¥
TTT¥¥¥¥########I++*HTIT+.-+*-+IITHH
TTTH¥¥¥¥######¥I!I*HH*+~+-++-~TTTIT
ITITTT¥¥######¥TTIH¥¥TI~~--++1HTT-+
-+~*TTH¥¥###H##TT*T¥H1++**-***HTTT*
-.-:tTT¥¥¥¥¥¥HH¥##:tIHHI++I-++ITHHTH*
-*I1H¥HH¥¥H¥¥¥#HTIHT**I+-*+TTITHT
-IITHHTTH¥HHH¥¥HI1THII+-.++.++TT+
*IITHTTHH¥¥THH¥HTI1I1-- +-+++*T.
--+1TTH¥H¥¥¥TTTTTTHH1I++** •
+*ITT¥¥#H¥T11I*THHH1I*--+1* ••
**TT¥¥¥¥THT**+IHHHHTT*- .---++- ••
*TT¥¥¥¥H¥HHTI1+1HTT1ITT-.++-**--.T*H¥##¥¥¥¥HT*-I*1H¥IIIH*+**
.--*1
HTH¥###¥¥T*T*--+.T¥¥HHHHT-*. -. ++*
H¥¥¥####¥¥¥HI*
HT¥¥#H¥¥H-**--.
-T
¥¥¥¥#####¥¥HII*.H¥H##¥¥¥¥**- - ++-*
¥¥¥########HTT+.1¥¥##¥##TI++
+.
¥¥######¥¥H***++TH#¥¥¥##HT+
--++
¥¥######¥¥H**+*¥¥¥#¥¥H###¥+
.*1*
#########HT11+1HH#HHHI¥###T+..
***
#########¥+*I11T¥H¥¥¥1T¥#¥HI*
.*1
#########HTT*¥¥¥¥###¥H*T##¥¥**+ .T*
#########¥¥THH¥¥¥¥¥#HTHT¥#¥H*
+.T1
##########¥¥##¥TTH¥TT1T-+T¥HT. ++*1
##########¥¥##¥HH1H***T-*!¥HT* 0+-1
########¥¥¥¥##HTT111-*T 1HTI* *--########¥####¥HIT111-.T- -IT1*+11##############¥ITT1 •• 11+ +1*11*+1*+
#############¥¥1I11 •• *+ -T1**II+-+########¥####¥##HHT+ ++. -I*+*ITTTT
########¥¥##H¥¥¥TH**.+I+ -*++**H¥¥T
###########¥HHI+T
•• *1
.+I.-H¥¥¥
###########¥1TI++.
*1II++T¥¥¥
############¥¥ITTHI
-+*IIHH###
00
11111111111111111111111122222222222
77778888888888999999999900000000001
67390123456789012345678901234567890
M.P.M. Estimated Map
VII ... 420
LEGEND
I
I
Residual percentage (%)
I
I
I
I
I
o
19
20
40
41
39
40
I
I
CHARACTER
ASSIGNED
I
I
I
I
BLANK
,
*I
00000000011111111112222222222333333
12345678901234567890123456789012345
,
1
2 ,
3
"
4
5
6
7
8
,
,
,
,
", ,
"
9
10
11
12
13
14 ,
15 "
16 ,
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
,"
,
,
, ,
"
"
"
,
5
,
I
"
,
"
, , ,,:II:" ,I,
" , ", ,
, ,
,
,
,
,
I"
,
,I
I
,
, ,
,
,.
"
,
"
"
I,
",
"
"
4
6
,
"
"
,
,
, ,
"""
""I,
,
" ,
1
2
3
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
3S
00000000011111111112222222222333333
12345678901234567890123456789012345
Fig.5
M.P.M. Estimated Residual Map
(TM data and AVHRR data)
VII
1
I
I
I
I
I
I
I