9310016,
i
ii
Abstract This thesis is on the researching of recognizing the roads in map image by computer. Based on the theory of Pattern Recognition, there is a method to be discussed, which can recognize and classify the different elements of map by color. First, the application background is introduced. Then there is some necessary analysis on the recognition object. In chapter 3, the method of classifying by color the different elements in map image is discussed in detail, which is mainly based on a theory in Pattern Recognition named Clustering. At last, the main models of programming are given. And some suggestion on problems to affect efficiency is proposed. iii
1....1 1.1... 1 1.2... 1 1.3... 2 2....3 2.1... 3 2.2... 4 3....4 3.1... 4 3.2... 7 3.3... 9 3.3.1... 9 3.3.2... 10 3.3.3... 11 4....13 4.1... 13 4.2... 14 5....14...16...17 iv
1. 1.1 GIS 1.2 Geographic Information System GIS 60 GIS GIS GIS 1
GIS GIS GIS 1.3 2
2. 2.1 3
2.2 (BMP) (pixel) m n m n (R) (G) (B) 255 0 256 3. 3.1 4
/ 5
6
3.2 RGB P = { p 11, p 12,..., p mn } pij i j m n RGB p ij = ( red ( pij), green ( pij), blue ( pij)) * * * * * * * * * 7
RGB mi dcolor(pi j,m i) pi j mi dcolor( p ij, mi) = Ni 1 mi = p Ni k= 1 k 2 w1 ( r( pij) r( mi)) + w2( g( pij) g( mi)) + w3( b( pij) b( mi)) 2 w1=4 w2=8 w3=1 8 8 3 3 1 0 3 3 1 1 0 3 3 1 0 2 8
3.3 RGB 3.3.1 a. b. c. (R,G,B 0) 9
3.3.2 P1 P8 P2 P0 P3 P4 P7 P6 P5 3 3 P0 10
a. P0(1)P1 3 5 7 P2 4 6 8 (2)P1 P8 5 P0 P1 P8 P0 P0 P0 P0 b. P0 8 c. 3.3.3 d1 d1 d2 11
0 1 2 3 0 0 d2 d2 G B 255 0 1 0 1 8 3 3 12 12
4. Borland Delphi for Windows 4.1 N Y N Y 13
4.2 Y Y N N N Y N Y 5. 14
RGB -- Fourier 15
16
[1] [2] [3] [4] [5] 95 6(7) [6] 93 27(6) [7] 96 [8] 93 [9] ART 94 [10] 17