核酸管外观缺陷检测(一)

1.1 应用示例思路

(1) 对核酸管图像进行灰度化、阈值分割和连通域分析;

(2) 筛选出待检测的区域,并对该区域进行变换校正;

(3) 进一步获取待检测的ROI区域,并根据几何特征和阈值条件,来对核酸管外观进行检测;

(4) 将检测结果写入txt文档。

1.2 应用示例代码

*统计核酸管外观正常的数组
P_Tuple:=[]
*统计核酸管外观缺陷的数组
N_Tuple:=[]
*判断文件是否存在
file_path:= './核酸管外观缺陷检测统计1.txt'
file_exists (file_path, file_exist)
if (file_exist==1)
    delete_file (file_path)
endif
open_file (file_path, 'output', FileHandle)
*获取文件路径列表
list_files ('./Test_img/', ['files','follow_links'], ImageFiles)
*文件筛选
tuple_regexp_select (ImageFiles, ['\\.(tif|tiff|gif|bmp|jpg|jpeg|jp2|png|pcx|pgm|ppm|pbm|xwd|ima|hobj)$','ignore_case'], ImageFiles)

for Index := 0 to |ImageFiles| - 1 by 1
    img_path:=ImageFiles[Index]
    *img_path:='./Test_img/Image_20230420091226458.bmp'
    *文件名拆分
    parse_filename (img_path, BaseName, Extension, Directory)
    *读取核酸管图片
    read_image (Image, img_path)
    *灰度化
    rgb1_to_gray (Image, GrayImage)
    *阈值分割
    threshold (GrayImage, Region, 30, 145)
    *填充孔洞
    fill_up (Region, RegionFillUp)
    *连通区域分析
    connection (RegionFillUp, ConnectedRegions)
    *获取面积
    area_center (ConnectedRegions, Area, Row, Column)
    *通过面积筛选区域
    select_shape (ConnectedRegions, SelectedRegions, 'area', 'and', 500000, 800000)
    *获取区域外接矩形
    smallest_rectangle2 (SelectedRegions, Row1, Column1, Phi, Length1, Length2)
    *创建变换模型
    vector_angle_to_rigid (Row1, Column1, abs(Phi), Row1, Column1, acos(0.0), HomMat2D)
    *进行区域变换
    affine_trans_region (SelectedRegions, RegionAffineTrans, HomMat2D, 'nearest_neighbor')
    *获取区域外接矩形(平行于坐标轴)
    smallest_rectangle1 (RegionAffineTrans, Row11, Column11, Row12, Column12)
    *裁剪区域
    clip_region (RegionAffineTrans, RegionClipped1, Row11, Column12-200, Row12, Column12)
    *获取区域外接矩形(平行于坐标轴)
    smallest_rectangle1 (RegionClipped1, Row21, Column21, Row22, Column22)
    *裁剪区域
    clip_region (RegionClipped1, RegionClipped2, int((Row21+Row22)/2.0)-20, Column12-200, int((Row21+Row22)/2.0)+20, Column12)
    *获取区域外接矩形(平行于坐标轴)
    smallest_rectangle1 (RegionClipped2, Row31, Column31, Row32, Column32) 
    distance:=Column32-Column31
    if(distance>=180)
        dev_clear_window ()
        get_image_size (Image, Width, Height)
        dev_open_window (0, 0, Width/2, Height/2, 'black', WindowHandle)
        dev_display (Image)
        *设置颜色
        dev_set_color ('green')
        *显示字符
        set_tposition (WindowHandle, 10, 5)  //设置文本光标1的位置
        write_string (WindowHandle, '核酸管外观正常!')
        set_tposition (WindowHandle, 50, 5)  //设置文本2光标的位置
        write_string (WindowHandle, ['核酸管外观正常的图片名:',BaseName])
        dev_close_window ()
        P_Tuple:=[P_Tuple,1]
        fwrite_string(FileHandle,[Index,'核酸管外观正常的图片名:',BaseName])
        fnew_line(FileHandle)    
    else
        dev_clear_window ()
        get_image_size (Image, Width, Height)
        dev_open_window (0, 0, Width/2, Height/2, 'black', WindowHandle)
        dev_display (Image)
        *设置颜色
        dev_set_color ('green')
        *显示字符
        set_tposition (WindowHandle, 10, 5)  //设置文本光标1的位置
        write_string (WindowHandle, '核酸管外观有缺陷!')
        set_tposition (WindowHandle, 50, 5)  //设置文本2光标的位置
        write_string (WindowHandle, ['核酸管外观缺陷的图片名:',BaseName])
        dev_close_window ()
        N_Tuple:=[N_Tuple,0]
        fwrite_string(FileHandle,[Index,'核酸管外观缺陷的图片名:',BaseName])
        fnew_line(FileHandle)    
    endif
endfor

tuple_length (P_Tuple, P_Length)
tuple_length (N_Tuple, N_Length)
Yield_Rate:= real(P_Length)/real(N_Length+P_Length)
fwrite_string(FileHandle,['核酸管良品率:',Yield_Rate])
fnew_line(FileHandle)
close_file (FileHandle)

1.3 结果展示

(1) 单张图片检测结果:

核酸管外观缺陷检测(一)_第1张图片

 (2) 部分图片检测结果:

核酸管外观缺陷检测(一)_第2张图片 

你可能感兴趣的:(《机器视觉应用》,机器视觉)