赞
踩
基于SVM先实现分类功能,对应不同缺陷类别完成分类;之后单独对应每一种具体缺陷来做缺陷位置定位。
- dev_update_off ()
- dev_close_window ()
- dev_open_window (0, 0, 400, 300, 'black', WindowHandle)
- dev_set_line_width (2)
-
- *创建支持向量机分类器
- create_class_svm (10, 'rbf', 0.02, 0.02, 4, 'one-versus-one', 'normalization', 10, SVMHandle)
-
- *确定图像和分类名称一一对应
- FileNames:=['无缺陷','凹点','划痕','锈斑污渍']
- for Classnumber := 0 to 3 by 1
- list_image_files (FileNames[Classnumber], 'default', [], ImageFiles1)
- for I := 0 to |ImageFiles1|-1 by 1
- *读取图像并筛选出各类特征
- read_image (Image, ImageFiles1[I])
- binary_threshold (Image, Region1, 'max_separability', 'light', UsedThreshold)
- fill_up (Region1, RegionFillUp)
- connection (RegionFillUp, Conn

Copyright © 2003-2013 www.wpsshop.cn 版权所有,并保留所有权利。