2 (62) 2 (654) E-mail wei.liu@changhong.com TP39.4 93 M. Albet Hanson Pinted Cicuit Boad PCB 936 D. Paul Eisne PCB PCB PCB (a)-(d) PCB PCB PCB 2 2(a) 2(b) PCB PCB PCB 5% [] 2(c) Automated optical Inspection AOI PCB AOI --
Diagnosys Tea dyne Phoenics AOI PCB PCB AOI PCB PCB PCB 2 x y L ƒ(x, ƒ(x, s (x, () [] + + f s ( x, s( x, f ( i x, j δ ( x i x, y j i j ƒ S (x, [ƒ(x,] M N M N 3dpi.8.mm.5mm ƒ S (x, L M N L ƒ(x, 3(a) 3dpi 256 PCB 3(b)-(d) -2-
92 5 29 49 52 87 8 29 44 59 84 4 3 45 56 8 4 25 43 5 85 5 23 42 53 R 2 3 2 3 3 2 3 3 2 3 4 3 88 88 84 82 87 85 83 9 86 82 8 78 87 9 84 86 83 77 85 86 84 82 85 84 86 B Colo Space RB CMY YCC/YUV Lab HSI H S I RB HSI 234 [2] B H { θ 36 θ B>.5[( R ) + ( R B)] θ accos 2 / [( R B) + ( R )( B)] 2 3 S [min( R,, B)] ( R + + B) I ( R + + 3 B ) R B [ ] H [ 36] S I [ ] 3 3. HSI H 4(a) 4(a) HSI H H 4(b) H 4(c) H 5 35 5 4(d) H 5 b(x, -3-
b ( x, y ) { f H f ( x, y ) [5 ( x, y ) [5 H, 35 ], 35 ] 6 5 5 4 3 5 2 5 4 2 5 5 2 25 3.2 5(a)-(d) PCB PCB 4 4. S 6(a) 3(a) HSI S S 6(b) S -4-
6(c) S 6(d) S s.9 6(e) 3(a) HSI I I 6(g) I S 6(h) I l.8 s.9 S 6 b(x, b( x, {.5 5 4.2 7(a)-(d) PCB PCB 2 3 25 2 5 5 4.2.4.6.8.8.6.4.2 5 [,.9] [,.8] S I (.9, ] (.8, ] S I 2 4 35 3 25 2 5 5 4.2.4.6.8-5-
5 5. 256 R B 255 8(a)-(d) RB 9(c) (f) (h) 9 (a) (e) (g) 9 (c) 9 (f) 9 (h) 95 72 [3] -6-
T Temp Temp T T ( n / n) ( n / n) 254 Temp2 n / n T + Temp3 254 ( ) ( n / n) T + Tnext ( Temp / Temp + Temp2 / Temp3) / 2 255 254 27 254 72 Tnext T Tnext < ε Tnext TTnext 72 3 Tnext Tnext PCB (4) 9 d (e) (g) b( x, { 3 2 [ Tnext, 254] [, Tnext) f ( x, 255 3 2 6 4 2 4 2 5 5 2 25 3 2 5 5 2 25 6 5 4 3 2 5 5 2 25-7-
5.2 (a) (c) PCB 6. PCB PCB PCB [] PCB [ ]22.4 [2] onzalez & Woods,. MATLAB,.,25.9 [3] 24,3 2 3 Reseach on the theshold segmentation algoithm about digital scanned image of cabon in PCB Liu wei Wang Weixing 2 Sichuan Changhong Component Technology Co.,LTD., Mianyang, PRC, (62) 2 School of Electonic Engineeing,Univesity of Electonic Science and Technology, Chengdu, PRC, (654) Abstact This pape povides the cuent status of detection technology of pinted cicuit boad and in aiming at the cuent limitation of detection technology, it innovatively poposed a new detection method based on digital scanned image.fo the new method, in this pape we analyzed many ey technologies about detection object segmentation of cabon in PCB, and many valuable eseached achievements wee got afte the veification of the expeiments Keywods: Pinted cicuit boad Detection Digital image 977-8-