基于Halcon学习的缺陷检测【一】detect_mura_blur.hdev

这个例子展示了如何在模糊图像中检测mura缺陷

---

总代码：

*频域+差分+空间域
dev_close_window ()
dev_update_off ()
Path := 'lcd/mura_defects_blur_'
read_image (Image, Path + '01')
get_image_size (Image, Width, Height)
dev_open_window_fit_size (0, 0, Width, Height, 640, 480, WindowHandle)
set_display_font (WindowHandle, 14, 'mono', 'true', 'false')
dev_set_draw ('margin')
dev_set_line_width (3)
dev_set_color ('red')
ScaleFactor := 0.4

*根据要提取的线条的最大宽度和对比度，计算线条的Sigma、Low和High参数。
calculate_lines_gauss_parameters (17, [25,3], Sigma, Low, High)
for f := 1 to 3 by 1
    *1采集图像
    read_image (Image, Path + f$'.2i')
    *转换为3通道图像
    decompose3 (Image, R, G, B)
    *2预处理之频域变换（建立背景+高斯滤波）
    * correct side illumination
    
    *正确的侧面照明
    *快速傅里叶变换
    rft_generic (B, ImageFFT, 'to_freq', 'none', 'complex', Width)
    *生成一个高斯滤波器
    gen_gauss_filter (ImageGauss, 100, 100, 0, 'n', 'rft', Width, Height)
    *在频域中用滤波器卷积图像。
    convol_fft (ImageFFT, ImageGauss, ImageConvol)
    *傅里叶逆变换
    rft_generic (ImageConvol, ImageFFT1, 'from_freq', 'none', 'byte', Width)
    *差分在空间域--lines_gauss
    sub_image (B, ImageFFT1, ImageSub, 2, 100)
    
    * perform the actual inspection
    *按给定因子缩放图像。
    zoom_image_factor (ImageSub, ImageZoomed, ScaleFactor, ScaleFactor, 'constant')
    * avoid border effects when using lines_gauss()
    *使用时lines_gauss()避免边界效果
    *获取图像的区域。
    get_domain (ImageZoomed, Domain)
    *腐蚀
    erosion_rectangle1 (Domain, RegionErosion, 7, 7)
    *抠图
    reduce_domain (ImageZoomed, RegionErosion, ImageReduced)
    *检测线条及其宽度。
    lines_gauss (ImageReduced, Lines, Sigma, Low, High, 'dark', 'true', 'gaussian', 'true')
    *进行仿射变换
    hom_mat2d_identity (HomMat2DIdentity)
    hom_mat2d_scale_local (HomMat2DIdentity, 1 / ScaleFactor, 1 / ScaleFactor, HomMat2DScale)
    affine_trans_contour_xld (Lines, Defects, HomMat2DScale)
    * 
    dev_display (Image)
    dev_display (Defects)
    if (f < 3)
        disp_continue_message (WindowHandle, 'black', 'true')
        stop ()
    endif
endfor

---

逐段分析：

*频域+差分+空间域
dev_close_window ()
dev_update_off ()
Path := 'lcd/mura_defects_blur_'
read_image (Image, Path + '01')
get_image_size (Image, Width, Height)
dev_open_window_fit_size (0, 0, Width, Height, 640, 480, WindowHandle)
set_display_font (WindowHandle, 14, 'mono', 'true', 'false')
dev_set_draw ('margin')
dev_set_line_width (3)
dev_set_color ('red')
ScaleFactor := 0.4

*根据要提取的线条的最大宽度和对比度，计算线条的Sigma、Low和High参数。
calculate_lines_gauss_parameters (17, [25,3], Sigma, Low, High)
for f := 1 to 3 by 1
    *1采集图像
    read_image (Image, Path + f$'.2i')

    *转换为3通道图像
    decompose3 (Image, R, G, B)

    *快速傅里叶变换
    rft_generic (B, ImageFFT, 'to_freq', 'none', 'complex', Width)

    *生成一个高斯滤波器
    gen_gauss_filter (ImageGauss, 100, 100, 0, 'n', 'rft', Width, Height)

    *在频域中用滤波器卷积图像。
    convol_fft (ImageFFT, ImageGauss, ImageConvol)

    *傅里叶逆变换
    rft_generic (ImageConvol, ImageFFT1, 'from_freq', 'none', 'byte', Width)

    *差分在空间域--lines_gauss
    sub_image (B, ImageFFT1, ImageSub, 2, 100)

    *按给定因子缩放图像。
    zoom_image_factor (ImageSub, ImageZoomed, ScaleFactor, ScaleFactor, 'constant')

    *使用时lines_gauss()避免边界效果
    *获取图像的区域。
    get_domain (ImageZoomed, Domain)

    *腐蚀
    erosion_rectangle1 (Domain, RegionErosion, 7, 7)

    *抠图
    reduce_domain (ImageZoomed, RegionErosion, ImageReduced)

    *检测线条及其宽度。
    lines_gauss (ImageReduced, Lines, Sigma, Low, High, 'dark', 'true', 'gaussian', 'true')

*进行仿射变换
    hom_mat2d_identity (HomMat2DIdentity)
    hom_mat2d_scale_local (HomMat2DIdentity, 1 / ScaleFactor, 1 / ScaleFactor, HomMat2DScale)
    affine_trans_contour_xld (Lines, Defects, HomMat2DScale)
    * 
    dev_display (Image)
    dev_display (Defects)
    if (f < 3)
        disp_continue_message (WindowHandle, 'black', 'true')
        stop ()
    endif
endfor

---

主要算子：

lines_gauss(Image : Lines : Sigma, Low, High, LightDark, ExtractWidth, LineModel, CompleteJunctions : 

功能：检测线条及其宽度

Image：输入图像

Lines：提取出的亚像素精度线条

Sigma：应用的高斯平滑的系数

Low：后滞阈值分割的低值

High：后滞阈值分割的高值

LightDark：提取图像中的亮色或者暗色线条

ExtractWidth：是否提取线条的宽度

LineModel：用于校正线条位置和宽度的线条模型

CompleteJunctions：是否添加能够提取的接合点