识别车牌号，哪个步骤不对呢，为什么识别的数字多几位

车牌号：粤BW39S5
class变量：['Y', 'O', 'F', '5', '5', '5', 'F', '5', '5', '5', '5', 'Y', '7', 'F', 'Y', 'S', '0', 'E', 'Q', 'H', 'E', '3', 'J', '4', 'I', '6', 'W', 'A', '4', '4', '0', '6', 'Y', 'L', '5', '5', '5', '5', 'N', '4', 'Y', 'A', '3', '3', 'S', '9', '1', '4', 'W', 'S', 'M', '5', '6', '5', '4', '4', '7', 'Q', '9', '9', 'E', 'W', 'B', '4', '9', 'N', 'W', '3', '9', 'I', '9', '9', 'W', 'S', '5', '4', 'O', 'E', 'Y', '6', 'S', 'W', 'W', '3', 'L', 'W', '6', '4', 'P', '4', '4', '6', '9', 'O', 'W', '9', '4', '5', '5', '1', '7', '5', 'B', '5', '5', 'Y', '4', '4', '3', '3', '6', 'S', '5', '4', '5', '5', '4', '9', '4', '4', '5', 'I', '7', 'R']

1. *读取图片
   
2. read_image(Image,'C:/Users/ff_pe/Desktop/图片/粤B.png')
   
3. *按三个图像转成三个颜色
   
4. *车牌定位矫正
   
5. decompose3(Image, Red, Green, Blue)
   
6. trans_from_rgb(Red, Green, Blue,Hue, Saturation, Intensity, 'hsv')
   
7. threshold (Saturation, Regions, 0, 190)
   
8. *现在是同一个连通域，需要断开
   
9. connection(Regions, ConnectedRegions)
   
10. *选中车牌号
    
11. select_shape (ConnectedRegions, SelectedRegions, 'area', 'and', 0, 458784)
    
12. *膨胀
    
13. dilation_rectangle1(Regions, RegionDilation, 30, 30)
    
14. *填空洞
    
15. fill_up(RegionDilation, RegionFillUp)
    
16. *车牌转正
    
17. orientation_region(RegionFillUp, Phi)
    
18. area_center(RegionFillUp, Area, Row, Column)
    
19. vector_angle_to_rigid(Row, Column, Phi, Row, Column, rad(180), HomMat2D)
    
20. affine_trans_image(Image, ImageAffineTrans, HomMat2D, 'constant', 'false')
    
21. affine_trans_region(RegionFillUp, RegionAffineTrans, HomMat2D, 'nearest_neighbor')
    
22. reduce_domain(ImageAffineTrans, RegionAffineTrans, ImageReduced)
    
23. 
    
24. *字符分割
    
25. rgb1_to_gray(ImageReduced, GrayImage)
    
26. invert_image(GrayImage, ImageInvert)
    
27. threshold (GrayImage, Regions1, 0, 87)
    
28. connection(Regions, ConnectedRegions)
    
29. select_shape (ConnectedRegions, SelectedRegions1, 'area', 'and', 0, 1.07161e+006)
    
30. 
    
31. read_ocr_class_mlp ('Industrial_0-9A-Z_NoRej.omc', OCRHandle1)
    
32. do_ocr_multi_class_mlp(SelectedRegions1, ImageInvert, OCRHandle1, Class, Confidence)
    

复制代码

图片在哪里？？

图图呢。

检查一下SelectedRegions1记录个数就行了

解决了这个问题
select_shape (SelectedRegions1, SelectedRegions1, 'area', 'and', 0, 5000)
改了这句

添加图片，就提示字数太大怎么办呢

给张图片看看