程式狂想曲

 
建立校正資料開始:
create_calib_data( : : CalibSetup, NumCameras, NumCalibObjects : CalibDataID)
  CalibSetup：校正模式
  NumCameras：相機數量
  NumCalibObjects：校正版數量
  CalibDataID：校正資料ID

適用  Area Camera, 使用 CCTV 鏡頭

CalPath := 'C:/Program Files/MVTec/HALCON-12.0/calib/自建/calplate_big.cpd' '校正版檔案路徑
StartCamPar := [0.016,0,0.0000074,0.0000074,320,240,640,480] '相機初始參數

Halcon 在3D 模組使用前需要將相機做參數校正
首先 需要設定 要使用的相機組成(CameraType)  與 相機與鏡頭的參數(CameraPAram)
 
CameraType 這邊參考 Halcon 的文件 選擇 合適的 Type

* Image Acquisition 01: Code generated by Image Acquisition 01
list_files ('D:/Study/鐵片方型尋找', ['files','follow_links'], ImageFiles)
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
    read_image (Image, ImageFiles[Index])
    rgb1_to_gray(Image, GrayImage)
    intensity(Image,GrayImage,Mean, Deviation)
    threshold(GrayImage,Region,0,Mean+Deviation*0.5)
    difference(GrayImage,Region,diffReg)
    fill_up(diffReg, RegionFillUp)
    reduce_domain(GrayImage, RegionFillUp, ImageReduced)
    threshold(ImageReduced,Region2,0,Mean)
    fill_up(Region2, RegionFillUp1)
    boundary(RegionFillUp1, RegionBorder, 'inner')
    connection(RegionBorder, ConnectedRegions)
    select_shape(ConnectedRegions,SelectedRegions,'contlength','and',190, 250)
    count_obj(SelectedRegions, Number)
    if (Number > 0)
        dev_display(GrayImage)
        for index := 1 to Number by 1
              select_obj(SelectedRegions, ObjectSelected, index)
              gen_contour_region_xld(ObjectSelected,Contours, 'border')
              shape_trans_xld(Contours,XLDTrans,'rectangle2')
              smallest_rectangle2_xld (XLDTrans, Row, Col, Phi, Len1, Len2)
              tuple_sin (Phi, Sin)
              tuple_cos (Phi, Cos)
              * 四点中心坐标 左，右，上，下
              L_Row := Row - (Len2 * Cos)
              L_Col := Col - (Len2 * Sin)
              B_Row := Row + (Len1 * Sin)
              B_Col := Col - (Len1 * Cos)
              T_Row := Row - (Len1 * Sin)
              T_Col := Col + (Len1 * Cos)
              R_Row := Row + (Len2 * Cos)
              R_Col := Col + (Len2 * Sin)
              gen_arrow_contour_xld (Arrow1, L_Row, L_Col, R_Row, R_Col, 1, 1)
              gen_arrow_contour_xld (Arrow2, B_Row, B_Col, T_Row, T_Col, 1, 1)
              gen_contour_polygon_xld (ContourL, [L_Row + 10 * Sin/2, L_Row - 10 * Sin/2], [L_Col - 10 * Cos/2, L_Col + 10 * Cos/2])
              gen_contour_polygon_xld (ContourR, [R_Row + 10 * Sin/2, R_Row - 10 * Sin/2], [R_Col - 10 * Cos/2, R_Col + 10 * Cos/2])
              gen_contour_polygon_xld (ContourT, [T_Row - 10 * Cos/2, T_Row + 10 * Cos/2], [T_Col - 10 * Sin/2, T_Col + 10 * Sin/2])
              gen_contour_polygon_xld (ContourB, [B_Row - 10 * Cos/2, B_Row + 10 * Cos/2], [B_Col - 10 * Sin/2, B_Col + 10 * Sin/2])
              dev_set_color ('blue')
              dev_display (Arrow1)
              dev_display (ContourL)
              dev_display (ContourR)
              dev_set_color ('green')
              dev_display (Arrow2)
              dev_display (ContourT)
              dev_display (ContourB)
              disp_message (3600, Len1 * 2, 'image', T_Row, T_Col, 'red', 'false')
              disp_message (3600, Len2 * 2, 'image', R_Row, R_Col, 'magenta', 'false')
             
             * dev_display(ObjectSelected)
              wait_seconds(1)
        endfor            
    endif    
endfor

閥值 threshold : 利用亮度值  做區域分割
 
auto_threshold : 自動閥值
參數 : sigma 

ㄧ、常見育處理影像方法:
1. mean_image or binomial_filter 消除雜訊
2.mediam_image 抑制小斑點或細線
3.smooth_image 平滑圖像

滤波（平滑、降噪）
增强
边缘锐化
纹理分析（去骨架、连通性）