, .
( ): 0.3-15 . - , 15-100 - , 100 - ( ). : 0.38-0.72 - ; 0.72-1.3 - ( ); 1.3 -3.0 - ; 7.0 - 15.0 - (). - .
3 7 ? . . , .
σR=F(N,V,n,no)A,4. V - , N - , , - .
, , . - , () (0.4 - 0.5). .
60- 20- - , . - , ( ). . - . - () .
20- . , , , , - , .
, . . , . , - , . -, , , -, . , , -, , .
. . ERDAS Imagine. . ERDAS Imagine.
|
|
|
- - (), (LANSAT, SPOT, , IRS, Terra .). , .
- - , .
. - , .
- .. , - -, , - , , , - , - - . - , . , . , - . ERDAS Imagine ( - -), , . , ( )
- , ( - ( ,))
Interpriter-Utiliteas
1.- Layer Stack .
: ,
: => -=>- =>
2.Subset-
-: , => ,
3. : . , .
- .
-: , () ( , => , .
|
|
|
(): -
:
-
-
: , -
1 - - - - . , ,, - - , , , - - , - - - - -, , - .
40 , 7
, , :
-
-
-
-
: - , , => , , .
:
: R=:4( ), G=3(), =2() - , - - . RGB- ( - -), - - - ( - ), - - ( - - ).
- : . : ( ), () . ERDAS Imagine .
- (1) ( ).
- n- (1) ( - ).
, :
;
.
, .
- ( - -). - , .
- => .
.
1. Classifier=>Signature Editor. Feature > View > Linked Cursor=> - => Link. . tm_86516.img. , - .
|
|
|
2. AOI > Tools => -=> - => - , => Feature > Masking > Image to Feature Space. -, , => Apply=> - , .
4. . , . . ERDAS Imagine.
- , .
- , - .
- n- -,
.
- -, n- .
- ( )
-, , .
, , , , - . : , , .
:
( - -), - - - ( - ), - - ( - - )
R, ( ):
, : , , , (: ).
. () - , n- X: 
X. , , . X. , .
|
|
|
=> , , . , . , ; . ; .
:
1.
AOI > Tools => -=> - => - , => Feature > Masking > Image to Feature Space. -, , => Apply=> - , .
2.
Classifier=>Signature Editor. Feature > View > Linked Cursor=> - => Link. . tm_86516.img. , - .
5. () . . , ERDAS Imagine.
() - -
-
-
: , -
1 - - - - . , ,, - - , , , - - , - - - - -, , - .
, 7
- (1) ( ).
- n- (1) ( - ).
, :
;
.
:
- - .
:
:
- - .=> =>
:
- -, , RGB- : (0.5 - 0.6 ), (0.6 - 0.7 ) () (0.7 - 0.8 0.8-0.9 ), RGB- .
RGB- ( - -), - - - ( - ), - - ( - - ).
|
|
|
:
-
, .
6. . ERDAS Imagine, .
C , , . - , .
- , . , , 3 4, - AOI- AOI > Tools. , , .
.
=> -=> AOI
( - , - -)
- => Feature > Statistics( ).
, , Feature > Masking > Feature Space to Image. Apply. , -,
, (), View > Image Alarm=> . , , .
, => Image Alarm Edit Parallelepiped Limits=> Set, Std. deviation => , .
=>
7. () . . . .
7 10 - , . , . , . , , , . , , .
:
-
-
: , -
8. , . ( ). ERDAS Imagine. . .
() - -
-
-
: , -
1 - - - - . , ,, - - , , , - - , - - - - -, , - .
- ( ), ISODATA( ).
- , .
: 1) ; 2) . , . .
ERDAS Imagine - : , ( convergence threshold).
: , , Skip factor (X Y=> , .
- (.5).
:
1. ( ).
=> => , . ( ) : => .
2. ,
.
ERDAS Imagine - Signature Means . . ( => , , , .
, , , , . .4.
, => 2-4 .
( .).
- => ,:- 2-, 3- 4-. . Subset Image DataPrep.
- tmJ360516.img,
- .
- : 2,3,4
- - - .
- Unsupervised Classification Classifier
- .
- .
- , , . 20 30. Color Scheme Options Approximate True Color, .
- Initialized Options , . .
-
- , ( , Raster Options, Clear Display =>)
- Raster > Attribute.
- Opacity. 0 Edit > Formula. Opacity 0 . , Color Opacity =1, , . . () Shift. Select None.
ISODATA. . . , ERDAS Imagine. .
ISODTA - .
7
:
1., ,
- Raster > Reeode .
- , , . +
- , Raster > Recode. New Value => Apply. .
2. =U ( Edit > Merge) .
- .
- .
- : Classify > Supervised. Parametric rule - .
.8() 2,3 ( 3,4 ). . 8 (), () - . () - , () - + .
() (), . () - . , .
:
- ( )
m3(2) ασ3(1)
m2(1) ασ1(2)
σ-
m-
1.
2. , σ ( )
3. =>
1- 1
3 1
:
- α
- ασ1(2) (2)-
- 2
-7 -
-6 =>
:
- (max )
- => =>
:
-
-
-, . , ,-
, - .
. , ERDAS Imagine.
, ERDAS Imagine - .
- , .
- - . - 1,2,3 2,3,4 , - 2,3 ( ), .
- Classifier Feature Space Thematic. :
1) (Thematic Input File);
2) Feature Space File - - ( 2,3 -);
3) Output Root Name 2-3 , .
- - Feature Space File
- - .
- .
- - .
- -, 2 Feature Space
- . AOI- - , - .
- , .
-
- , .
- , .
-- Feature >Objects. -, Plot Ellipse Plot Means
- - (.9). Plot Ellipse Plot Rectangle .
:
- , => , , .
- . , (). .
, :
- . , .
: , , . , , . , , .
, . . ( ) ERDAS Imagine.
, () .
( ) , , , . , , .
. ( , , .) . , , .
, AOI > Tools. & . +. - , , . Feature > Masking > Image to Feature Space. -, , Apply. - , . , - , . .. , , . - , .
. - , . , , 3 4, - AOI- AOI > Tools. , , - AOI . , . , , - . - , - -.
- . Feature > Statistics .
, , Feature > Masking > Feature Space to Image. Apply. , -, .
-, , , ( ).
, (), View > Image Alarm. . , , .
, , , Image Alarm Edit Parallelepiped Limits. Set, Std. deviation .
ERDAS Imagine. , . . , .
, () .
( ) , , , . , , .
Supervised Classification Classifier. - (Non-parametric Rule) (Parametric Rule). , ( Parallelepiped Non-parametric Rule), , - ( Feature Space Non-parametric Rule).
. , , :
(5.1)
Ω j () - xj, . =2, =3 . 5.1.
. 5.1 Ω1(l)=Ω1(2)+Ω1(3), A1 2 3 2. Ω2(2)=Ω2(3), 1
.5 . ().
. .10. , . . , . . .
. - - 9 ) .
{ Minimum Distance Parametric Rule) . . - , .
, . , () .
, , :
. , , . .10, , . . , , , . . ( ), . , , .
( Maksimum Likelihood Parametric Rule) . Ωk xi, :
p(Ωk/ xi)=(P(Ωk) p(xi /Ωk))/ (∑P(Ωj) p(xi /Ωj))
, . () - . , . . ( - ).
.
, 1 2 , 0. 0 1, - 2.
, , .
Supervised Classification Unclassified Rule. , , , , , dki(x) i , , , , - .
, , , , , . 20. 3, . , , : . , , .
10.
, - - , (Overlap Rule). , , . ( . ), . Use Probabilities.
( Order ). (, , Feature Space) , . , 2 5 2.
. . . .
, , . , , . k- Pk(x) , fk(x).
. . ,
1) ( );
2) X.
, X. X Ω 1 Ω 2, . Ω.
, : f1(x) f2(x) (.8.1), P(Ω1) (Ω2) . - , , , .
0, : (-∞,0] Ω1, (,∞) - Ω2. ξ, . : ξ, , f1(x), Ω 1 - H1 2: ξ f2(x), Ω.
.8.1. .
, X, 4 .
1. 1 .
2. 2, H1.
3. 2, .
4. 1 2.
1 f1(x) (-∞,0], 2 - f1(x) (,∞), 3 - f2(x) (,∞), 4 - f2(x) (-∞,]- f1(x) f2(x) 2 4 - . , 2, (), , 4, - (|3). , , , - . H 2 , 4.
Ω1 . , . , . , , , . , , , , . , , , .
. . .52 .
. , . , ERDAS Imagine. .
Ω1...,ΩK f1(x),...,fK(x) P(Ω1),...,P(Ωk). , . , S0, () . , () X;.
, Xj k- P(Ωk) fk(Ωj), , , (Ω)(Ω). i - xi -
k- ,
, Xi k- ,
(8.10) . i .
(Ω), k- , p(Ωk/Xi) - k-ro . .
P(Ωk)p(x/Qk)=P(Ωk)fk(x) ( P(Ωk)p(x/Ωk) ), (8.10) , s (8.6)
, , Lks(x)>λ (Lks()>1 C12==c21=l; , , . , , p(Ωk /x)> p(Ωs /x) , , (Ω)(/Ωk)>(Ωs)(/Ωs)) k,s. - .
, 0 1, . - . , .
, . , , (/Ωs). Pi(Ωs)=Sj.i(A), i - , a Si-1(Ωs) - Ωs, i-1- , [28].
:
Feature > Objects. , . , .
