{"id":57167,"date":"2025-08-15T08:20:05","date_gmt":"2025-08-15T00:20:05","guid":{"rendered":"https:\/\/www.wsisp.com\/helps\/57167.html"},"modified":"2025-08-15T08:20:05","modified_gmt":"2025-08-15T00:20:05","slug":"%e5%9f%ba%e4%ba%8eopencv%e7%9a%84%e5%ae%9e%e6%97%b6%e7%be%8e%e9%a2%9c%e6%8a%80%e6%9c%af%ef%bc%9a%e4%bb%8e%e4%bc%a0%e7%bb%9f%e7%ae%97%e6%b3%95%e5%88%b0%e6%b7%b1%e5%ba%a6%e5%ad%a6%e4%b9%a0%e8%9e%8d","status":"publish","type":"post","link":"https:\/\/www.wsisp.com\/helps\/57167.html","title":{"rendered":"\u57fa\u4e8eOpenCV\u7684\u5b9e\u65f6\u7f8e\u989c\u6280\u672f\uff1a\u4ece\u4f20\u7edf\u7b97\u6cd5\u5230\u6df1\u5ea6\u5b66\u4e60\u878d\u5408\u5b9e\u73b0"},"content":{"rendered":"

\u5f15\u8a00：\u7f8e\u989c\u6280\u672f\u7684\u73b0\u72b6\u4e0e\u6311\u6218<\/h3>\n

\u5728\u793e\u4ea4\u5a92\u4f53\u4e0e\u89c6\u9891\u901a\u4fe1\u666e\u53ca\u7684\u4eca\u5929，\u5b9e\u65f6\u7f8e\u989c\u5df2\u6210\u4e3a\u79fb\u52a8\u5e94\u7528\u4e0e\u76f4\u64ad\u5e73\u53f0\u7684\u6838\u5fc3\u529f\u80fd\u3002OpenCV\u4f5c\u4e3a\u5f00\u6e90\u8ba1\u7b97\u673a\u89c6\u89c9\u5e93，\u51ed\u501f\u5176\u9ad8\u6548\u7684\u56fe\u50cf\u5904\u7406\u63a5\u53e3\u4e0e\u8de8\u5e73\u53f0\u7279\u6027，\u6210\u4e3a\u5b9e\u73b0\u7f8e\u989c\u529f\u80fd\u7684\u7406\u60f3\u9009\u62e9\u3002\u672c\u6587\u5c06\u7cfb\u7edf\u4ecb\u7ecd\u57fa\u4e8eOpenCV\u7684\u7f8e\u989c\u6280\u672f\u5b9e\u73b0\u65b9\u6848，\u6db5\u76d6\u4f20\u7edf\u6ee4\u6ce2\u7b97\u6cd5\u3001\u6df1\u5ea6\u5b66\u4e60\u6df7\u5408\u6a21\u578b\u53ca\u6027\u80fd\u4f18\u5316\u7b56\u7565，\u5e2e\u52a9\u5f00\u53d1\u8005\u6784\u5efa\u517c\u987e\u81ea\u7136\u5ea6\u4e0e\u5b9e\u65f6\u6027\u7684\u7f8e\u989c\u7cfb\u7edf\u3002<\/p>\n

\u4e00\u3001\u73af\u5883\u51c6\u5907\u4e0e\u6838\u5fc3\u4f9d\u8d56<\/h3>\n

1.1 \u5f00\u53d1\u73af\u5883\u914d\u7f6e<\/h4>\n

# \u57fa\u7840\u4f9d\u8d56\u5b89\u88c5<\/span>
\npip install<\/span> opencv-python==<\/span>4.8<\/span>.0 numpy<\/span>==<\/span>1.26<\/span>.0 dlib<\/span>==<\/span>19.24<\/span>.2
\n# \u5982\u9700\u6df1\u5ea6\u5b66\u4e60\u6a21\u5757（\u53ef\u9009）<\/span>
\npip install<\/span> opencv-contrib-python openvino-dev<\/p>\n

1.2 \u5173\u952e\u5e93\u529f\u80fd\u8bf4\u660e<\/h4>\n\n\u5e93\u540d\u79f0\u6838\u5fc3\u4f5c\u7528\u7248\u672c\u8981\u6c42<\/tr>\n\n\n\n\n\n
opencv-python<\/td>\n\u63d0\u4f9b\u57fa\u7840\u56fe\u50cf\u5904\u7406\u4e0e\u6ee4\u6ce2\u51fd\u6570<\/td>\n\u22654.2.0<\/td>\n<\/tr>\n
dlib<\/td>\n\u5b9e\u73b0\u4eba\u8138\u7279\u5f81\u70b9\u68c0\u6d4b（68\u70b9\/35\u70b9）<\/td>\n\u226519.20.0<\/td>\n<\/tr>\n
numpy<\/td>\n\u77e9\u9635\u8fd0\u7b97\u52a0\u901f<\/td>\n\u22651.21.0<\/td>\n<\/tr>\n
openvino-dev<\/td>\n\u4f18\u5316\u6df1\u5ea6\u5b66\u4e60\u6a21\u578b\u63a8\u7406（\u53ef\u9009）<\/td>\n\u22652024.0<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

\u4e8c\u3001\u4f20\u7edf\u7f8e\u989c\u7b97\u6cd5\u539f\u7406\u4e0e\u5b9e\u73b0<\/h3>\n

2.1 \u53cc\u8fb9\u6ee4\u6ce2\u78e8\u76ae\u6280\u672f<\/h4>\n

\u53cc\u8fb9\u6ee4\u6ce2\u662f\u7f8e\u989c\u7684\u6838\u5fc3\u7b97\u6cd5，\u901a\u8fc7\u540c\u65f6\u8003\u8651\u7a7a\u95f4\u90bb\u8fd1\u5ea6（\u9ad8\u65af\u6838）\u4e0e\u50cf\u7d20\u76f8\u4f3c\u5ea6（\u7070\u5ea6\u5dee\u6743\u91cd），\u5b9e\u73b0\u201c\u4fdd\u8fb9\u964d\u566a\u201d\u6548\u679c\u3002\u5176\u6570\u5b66\u8868\u8fbe\u5f0f\u4e3a： <\/p>\n

W<\/p>\n

(<\/p>\n

i<\/p>\n

,<\/p>\n

j<\/p>\n

,<\/p>\n

k<\/p>\n

,<\/p>\n

l<\/p>\n

)<\/p>\n

=<\/p>\n

exp<\/p>\n

\u2061<\/p>\n

(<\/p>\n

\u2212<\/p>\n

(<\/p>\n

i<\/p>\n

\u2212<\/p>\n

k<\/p>\n

)<\/p>\n

2<\/p>\n

+<\/p>\n

(<\/p>\n

j<\/p>\n

\u2212<\/p>\n

l<\/p>\n

)<\/p>\n

2<\/p>\n

2<\/p>\n

\u03c3<\/p>\n

d<\/p>\n

2<\/p>\n

\u2212<\/p>\n

\u2225<\/p>\n

I<\/p>\n

(<\/p>\n

i<\/p>\n

,<\/p>\n

j<\/p>\n

)<\/p>\n

\u2212<\/p>\n

I<\/p>\n

(<\/p>\n

k<\/p>\n

,<\/p>\n

l<\/p>\n

)<\/p>\n

\u2225<\/p>\n

2<\/p>\n

2<\/p>\n

\u03c3<\/p>\n

r<\/p>\n

2<\/p>\n

)<\/p>\n

W(i,j,k,l) = \\\\exp\\\\left(-\\\\frac{(i-k)^2+(j-l)^2}{2\\\\sigma_d^2}-\\\\frac{\\\\|I(i,j)-I(k,l)\\\\|^2}{2\\\\sigma_r^2}\\\\right)<\/p>\n

<\/span><\/span>W<\/span>(<\/span>i<\/span>,<\/span><\/span>j<\/span>,<\/span><\/span>k<\/span>,<\/span><\/span>l<\/span>)<\/span><\/span>=<\/span><\/span><\/span><\/span>exp<\/span><\/span>(<\/span><\/span>\u2212<\/span><\/span><\/span>2<\/span>\u03c3<\/span><\/span>d<\/span><\/span><\/span><\/span>2<\/span><\/span><\/span><\/span>\u200b<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span>(<\/span>i<\/span><\/span>\u2212<\/span><\/span>k<\/span>)<\/span><\/span>2<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span>+<\/span><\/span>(<\/span>j<\/span><\/span>\u2212<\/span><\/span>l<\/span>)<\/span><\/span>2<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span>\u200b<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span>\u2212<\/span><\/span><\/span><\/span>2<\/span>\u03c3<\/span><\/span>r<\/span><\/span><\/span><\/span>2<\/span><\/span><\/span><\/span>\u200b<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span>\u2225<\/span>I<\/span>(<\/span>i<\/span>,<\/span><\/span>j<\/span>)<\/span><\/span>\u2212<\/span><\/span>I<\/span>(<\/span>k<\/span>,<\/span><\/span>l<\/span>)<\/span>\u2225<\/span><\/span>2<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span>\u200b<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span>)<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span> \u5176\u4e2d <\/p>\n

\u03c3<\/p>\n

d<\/p>\n

\\\\sigma_d<\/p>\n

<\/span><\/span>\u03c3<\/span><\/span>d<\/span><\/span><\/span><\/span>\u200b<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span> \u63a7\u5236\u7a7a\u95f4\u5e73\u6ed1\u8303\u56f4，<\/p>\n

\u03c3<\/p>\n

r<\/p>\n

\\\\sigma_r<\/p>\n

<\/span><\/span>\u03c3<\/span><\/span>r<\/span><\/span><\/span><\/span>\u200b<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span> \u63a7\u5236\u50cf\u7d20\u76f8\u4f3c\u5ea6\u9608\u503c\u3002<\/p>\n

Python\u5b9e\u73b0\u4ee3\u7801：<\/p>\n

def<\/span> bilateral_filter_skin_smoothing<\/span>(<\/span>image,<\/span> sigma_color=<\/span>30<\/span>,<\/span> sigma_space=<\/span>15<\/span>)<\/span>:<\/span>
\n """
\n \u53cc\u8fb9\u6ee4\u6ce2\u78e8\u76ae\u5b9e\u73b0
\n :param image: BGR\u683c\u5f0f\u8f93\u5165\u56fe\u50cf
\n :param sigma_color: \u989c\u8272\u7a7a\u95f4\u6807\u51c6\u5dee（\u63a8\u835030-75）
\n :param sigma_space: \u5750\u6807\u7a7a\u95f4\u6807\u51c6\u5dee（\u63a8\u835015-50）
\n :return: \u78e8\u76ae\u540e\u56fe\u50cf
\n """<\/span>
\n # \u8f6c\u6362\u4e3a\u6d6e\u70b9\u578b\u907f\u514d\u6ea2\u51fa<\/span>
\n img_float =<\/span> image.<\/span>astype(<\/span>np.<\/span>float32)<\/span>
\n # \u5e94\u7528\u53cc\u8fb9\u6ee4\u6ce2<\/span>
\n smoothed =<\/span> cv2.<\/span>bilateralFilter(<\/span>
\n img_float,<\/span> d=<\/span>0<\/span>,<\/span> sigmaColor=<\/span>sigma_color,<\/span> sigmaSpace=<\/span>sigma_space
\n )<\/span>
\n return<\/span> smoothed.<\/span>astype(<\/span>np.<\/span>uint8)<\/span><\/p>\n

2.2 \u80a4\u8272\u68c0\u6d4b\u4e0e\u533a\u57df\u4f18\u5316<\/h4>\n

\u4e3a\u907f\u514d\u5bf9\u5934\u53d1\u3001\u80cc\u666f\u7b49\u975e\u76ae\u80a4\u533a\u57df\u8fc7\u5ea6\u5904\u7406，\u9700\u901a\u8fc7\u80a4\u8272\u63a9\u7801\u7cbe\u51c6\u5b9a\u4f4d\u7f8e\u989c\u533a\u57df\u3002HSV\u8272\u5f69\u7a7a\u95f4\u4e0b，\u4e9a\u6d32\u4eba\u80a4\u8272\u8303\u56f4\u901a\u5e38\u4e3a：<\/p>\n

    \n
  • \u4e0b\u754c：[0, 20, 70]（H: 0-20\u00b0, S: 20-255, V: 70-255）<\/li>\n
  • \u4e0a\u754c：[20, 255, 255]<\/li>\n<\/ul>\n

    \u80a4\u8272\u63a9\u7801\u751f\u6210\u4ee3\u7801：<\/p>\n

    def<\/span> create_skin_mask<\/span>(<\/span>image)<\/span>:<\/span>
    \n """\u751f\u6210\u80a4\u8272\u533a\u57df\u63a9\u7801"""<\/span>
    \n hsv =<\/span> cv2.<\/span>cvtColor(<\/span>image,<\/span> cv2.<\/span>COLOR_BGR2HSV)<\/span>
    \n lower_skin =<\/span> np.<\/span>array(<\/span>[<\/span>0<\/span>,<\/span> 20<\/span>,<\/span> 70<\/span>]<\/span>,<\/span> dtype=<\/span>np.<\/span>uint8)<\/span>
    \n upper_skin =<\/span> np.<\/span>array(<\/span>[<\/span>20<\/span>,<\/span> 255<\/span>,<\/span> 255<\/span>]<\/span>,<\/span> dtype=<\/span>np.<\/span>uint8)<\/span>
    \n mask =<\/span> cv2.<\/span>inRange(<\/span>hsv,<\/span> lower_skin,<\/span> upper_skin)<\/span>
    \n # \u5f62\u6001\u5b66\u64cd\u4f5c\u4f18\u5316\u63a9\u7801（\u53bb\u9664\u566a\u58f0）<\/span>
    \n kernel =<\/span> cv2.<\/span>getStructuringElement(<\/span>cv2.<\/span>MORPH_ELLIPSE,<\/span> (<\/span>5<\/span>,<\/span> 5<\/span>)<\/span>)<\/span>
    \n mask =<\/span> cv2.<\/span>morphologyEx(<\/span>mask,<\/span> cv2.<\/span>MORPH_CLOSE,<\/span> kernel)<\/span>
    \n return<\/span> mask<\/p>\n

    2.3 \u591a\u7ea7\u4e2d\u503c\u6ee4\u6ce2\u53bb\u566a<\/h4>\n

    \u9488\u5bf9\u9762\u90e8\u6591\u70b9\u4e0e\u6912\u76d0\u566a\u58f0，\u8fed\u4ee3\u591a\u7ea7\u4e2d\u503c\u6ee4\u6ce2\u6548\u679c\u4f18\u4e8e\u4f20\u7edf\u5747\u503c\u6ee4\u6ce2\u3002\u901a\u8fc73\u00d73\u30015\u00d75\u7a97\u53e3\u4ea4\u66ff\u5904\u7406，\u53ef\u5728\u4fdd\u7559\u7eb9\u7406\u7ec6\u8282\u7684\u540c\u65f6\u53bb\u9664\u5f02\u5e38\u50cf\u7d20：<\/p>\n

    def<\/span> multi_stage_median_filter<\/span>(<\/span>image,<\/span> iterations=<\/span>2<\/span>)<\/span>:<\/span>
    \n """\u591a\u7ea7\u4e2d\u503c\u6ee4\u6ce2\u53bb\u566a"""<\/span>
    \n result =<\/span> image.<\/span>copy(<\/span>)<\/span>
    \n for<\/span> _ in<\/span> range<\/span>(<\/span>iterations)<\/span>:<\/span>
    \n result =<\/span> cv2.<\/span>medianBlur(<\/span>result,<\/span> ksize=<\/span>3<\/span>)<\/span> # 3\u00d73\u7a97\u53e3<\/span>
    \n result =<\/span> cv2.<\/span>medianBlur(<\/span>result,<\/span> ksize=<\/span>5<\/span>)<\/span> # 5\u00d75\u7a97\u53e3<\/span>
    \n return<\/span> result<\/p>\n

    \u4e09\u3001\u6df1\u5ea6\u5b66\u4e60\u6df7\u5408\u7f8e\u989c\u6846\u67b6<\/h3>\n

    3.1 \u57fa\u4e8eG-API\u7684\u7ba1\u9053\u6784\u5efa<\/h4>\n

    OpenCV 4.2+\u5f15\u5165\u7684G-API（\u56fe\u5f62\u5904\u7406API）\u652f\u6301\u6df1\u5ea6\u5b66\u4e60\u6a21\u578b\u4e0e\u4f20\u7edf\u7b97\u6cd5\u7684\u6d41\u6c34\u7ebf\u878d\u5408\u3002\u4ee5\u201c\u4eba\u8138\u68c0\u6d4b\u2192\u7279\u5f81\u70b9\u5b9a\u4f4d\u2192\u533a\u57df\u6ee4\u6ce2\u201d\u4e3a\u4f8b，\u6838\u5fc3\u6d41\u7a0b\u5982\u4e0b：<\/p>\n

  • \u4eba\u8138\u68c0\u6d4b：\u4f7f\u7528face-detection-adas-0001\u6a21\u578b（SSD\u67b6\u6784）\u751f\u6210\u4eba\u8138\u8fb9\u754c\u6846<\/li>\n
  • \u7279\u5f81\u70b9\u63d0\u53d6：\u901a\u8fc7facial-landmarks-35-adas-0002\u83b7\u53d635\u4e2a\u5173\u952e\u70b9（\u542b\u773c\u775b\u3001\u5634\u90e8\u7b49）<\/li>\n
  • \u8499\u7248\u751f\u6210：\u57fa\u4e8e\u5173\u952e\u70b9\u751f\u6210\u4e09\u7c7b\u8499\u7248（\u80cc\u666f\u8499\u7248b\u3001\u9510\u5316\u533a\u57df\u8499\u7248p\u3001\u6a21\u7cca\u533a\u57df\u8499\u7248s）<\/li>\n
  • \u878d\u5408\u8ba1\u7b97：\n

    O<\/p>\n

    =<\/p>\n

    b<\/p>\n

    \u2217<\/p>\n

    I<\/p>\n

    +<\/p>\n

    p<\/p>\n

    \u2217<\/p>\n

    U<\/p>\n

    +<\/p>\n

    s<\/p>\n

    \u2217<\/p>\n

    L<\/p>\n

    O = b*I + p*U + s*L<\/p>\n

    <\/span><\/span>O<\/span><\/span>=<\/span><\/span><\/span><\/span>b<\/span><\/span>\u2217<\/span><\/span><\/span><\/span>I<\/span><\/span>+<\/span><\/span><\/span><\/span>p<\/span><\/span>\u2217<\/span><\/span><\/span><\/span>U<\/span><\/span>+<\/span><\/span><\/span><\/span>s<\/span><\/span>\u2217<\/span><\/span><\/span><\/span>L<\/span><\/span><\/span><\/span><\/span>（<\/p>\n

    U<\/p>\n

    U<\/p>\n

    <\/span><\/span>U<\/span><\/span><\/span><\/span><\/span>\u4e3a\u9510\u5316\u56fe\u50cf，<\/p>\n

    L<\/p>\n

    L<\/p>\n

    <\/span><\/span>L<\/span><\/span><\/span><\/span><\/span>\u4e3a\u6a21\u7cca\u56fe\u50cf）<\/li>\n

    G-API\u7ba1\u9053\u6838\u5fc3\u4ee3\u7801：<\/p>\n

    # \u58f0\u660eDNN\u62d3\u6251（\u7b80\u5316\u7248）<\/span>
    \nG_API_NET(<\/span>FaceDetector,<\/span> <<\/span>cv:<\/span>:<\/span>GMat(<\/span>cv:<\/span>:<\/span>GMat)<\/span>><\/span>,<\/span> "face_detector"<\/span>)<\/span>
    \nG_API_NET(<\/span>LandmDetector,<\/span> <<\/span>cv:<\/span>:<\/span>GMat(<\/span>cv:<\/span>:<\/span>GMat)<\/span>><\/span>,<\/span> "landm_detector"<\/span>)<\/span><\/p>\n

    # \u6784\u5efa\u5904\u7406\u7ba1\u9053<\/span>
    \npipeline =<\/span> cv2.<\/span>gapi.<\/span>GComputation(<\/span>[<\/span>=<\/span>]<\/span>(<\/span>)<\/span> {<\/span>
    \n gimgIn =<\/span> cv2.<\/span>gapi.<\/span>GMat(<\/span>)<\/span>
    \n # \u4eba\u8138\u68c0\u6d4b\u4e0e\u7279\u5f81\u70b9\u63d0\u53d6<\/span>
    \n faceOut =<\/span> cv2.<\/span>gapi.<\/span>infer<<\/span>FaceDetector><\/span>(<\/span>gimgIn)<\/span>
    \n landmOut =<\/span> cv2.<\/span>gapi.<\/span>infer<<\/span>LandmDetector><\/span>(<\/span>faceOut)<\/span>
    \n # \u751f\u6210\u8499\u7248\u4e0e\u878d\u5408<\/span>
    \n garElsConts,<\/span> garFaceConts =<\/span> custom:<\/span>:<\/span>GGetContours:<\/span>:<\/span>on(<\/span>landmOut)<\/span>
    \n mskSharp =<\/span> custom:<\/span>:<\/span>GFillPolyGContours:<\/span>:<\/span>on(<\/span>gimgIn,<\/span> garElsConts)<\/span> # \u9510\u5316\u533a\u57df<\/span>
    \n mskBlur =<\/span> custom:<\/span>:<\/span>GFillPolyGContours:<\/span>:<\/span>on(<\/span>gimgIn,<\/span> garFaceConts)<\/span> # \u6a21\u7cca\u533a\u57df<\/span>
    \n # \u9ad8\u65af\u6a21\u7cca\u4e0e\u878d\u5408<\/span>
    \n mskSharpG =<\/span> cv2.<\/span>gapi.<\/span>gaussianBlur(<\/span>mskSharp,<\/span> (<\/span>7<\/span>,<\/span>7<\/span>)<\/span>,<\/span> 1.5<\/span>)<\/span>
    \n mskBlurG =<\/span> cv2.<\/span>gapi.<\/span>gaussianBlur(<\/span>mskBlur,<\/span> (<\/span>15<\/span>,<\/span>15<\/span>)<\/span>,<\/span> 3.0<\/span>)<\/span>
    \n return<\/span> mskSharpG *<\/span> gimgIn +<\/span> mskBlurG *<\/span> cv2.<\/span>gapi.<\/span>bilateralFilter(<\/span>gimgIn,<\/span> 9<\/span>,<\/span> 75<\/span>,<\/span> 75<\/span>)<\/span>
    \n}<\/span>)<\/span><\/p>\n

    3.2 YOLOv8+CodeFormer\u6df7\u5408\u6a21\u578b<\/h4>\n

    \u5bf9\u4e8e\u9ad8\u7cbe\u5ea6\u9700\u6c42，\u53ef\u7ed3\u5408\u76ee\u6807\u68c0\u6d4b\u4e0e\u7ec6\u8282\u4fee\u590d\u6a21\u578b：<\/p>\n

      \n
    • YOLOv8-Face：\u5b9e\u65f6\u68c0\u6d4b\u4eba\u8138\u533a\u57df（640\u00d7640\u8f93\u5165\u4e0bFPS\u8fbe30+）<\/li>\n
    • CodeFormer：\u57fa\u4e8eTransformer\u7684\u9762\u90e8\u7ec6\u8282\u4fee\u590d，\u89e3\u51b3\u8fc7\u5ea6\u6a21\u7cca\u5bfc\u81f4\u7684\u201c\u5851\u6599\u611f\u201d<\/li>\n
    • \u878d\u5408\u7b56\u7565：\u901a\u8fc7FaceParsing\u751f\u6210\u76ae\u80a4\u63a9\u7801，\u4ec5\u5bf9\u76ae\u80a4\u533a\u57df\u5e94\u7528\u7f8e\u989c，\u4fdd\u7559\u7709\u6bdb\u3001\u776b\u6bdb\u7b49\u7ec6\u8282<\/li>\n<\/ul>\n

      \u6a21\u578b\u63a8\u7406\u6d41\u7a0b：<\/p>\n

      # \u7b80\u5316\u793a\u4f8b：YOLOv8\u68c0\u6d4b + CodeFormer\u4f18\u5316<\/span>
      \nfrom<\/span> ultralytics import<\/span> YOLO
      \nfrom<\/span> codeformer import<\/span> CodeFormerInference<\/p>\n

      yolo_model =<\/span> YOLO(<\/span>"yolov8n-face.pt"<\/span>)<\/span>
      \ncodeformer =<\/span> CodeFormerInference(<\/span>"codeformer.pth"<\/span>)<\/span><\/p>\n

      def<\/span> deep_beautify<\/span>(<\/span>image)<\/span>:<\/span>
      \n # 1. \u4eba\u8138\u68c0\u6d4b<\/span>
      \n results =<\/span> yolo_model(<\/span>image)<\/span>[<\/span>0<\/span>]<\/span>
      \n for<\/span> det in<\/span> results.<\/span>boxes:<\/span>
      \n x1,<\/span> y1,<\/span> x2,<\/span> y2 =<\/span> map<\/span>(<\/span>int<\/span>,<\/span> det.<\/span>xyxy[<\/span>0<\/span>]<\/span>)<\/span>
      \n face =<\/span> image[<\/span>y1:<\/span>y2,<\/span> x1:<\/span>x2]<\/span>
      \n # 2. CodeFormer\u4f18\u5316<\/span>
      \n restored_face =<\/span> codeformer.<\/span>infer(<\/span>face,<\/span> fidelity=<\/span>0.7<\/span>)<\/span> # 0.7\u4e3a\u81ea\u7136\u5ea6-\u4fee\u590d\u5f3a\u5ea6\u5e73\u8861\u503c<\/span>
      \n # 3. \u8499\u7248\u878d\u5408<\/span>
      \n mask =<\/span> create_skin_mask(<\/span>restored_face)<\/span>
      \n image[<\/span>y1:<\/span>y2,<\/span> x1:<\/span>x2]<\/span> =<\/span> cv2.<\/span>bitwise_and(<\/span>restored_face,<\/span> restored_face,<\/span> mask=<\/span>mask)<\/span>
      \n return<\/span> image<\/p>\n

      \u56db\u3001\u5b8c\u6574\u7f8e\u989c\u7cfb\u7edf\u5b9e\u73b0<\/h3>\n

      4.1 \u529f\u80fd\u6574\u5408\u6d41\u7a0b\u56fe<\/h4>\n

      graph TD
      \n A[\u8f93\u5165\u56fe\u50cf] –> B[\u4eba\u8138\u68c0\u6d4b<br>Haar\u7ea7\u8054\/YOLOv8]
      \n B –> C{\u662f\u5426\u68c0\u6d4b\u5230\u4eba\u8138?}
      \n C –>|\u5426| D[\u76f4\u63a5\u8f93\u51fa\u539f\u56fe]
      \n C –>|\u662f| E[\u7279\u5f81\u70b9\u63d0\u53d6<br>68\u70b9\/35\u70b9]
      \n E –> F[\u751f\u6210\u80a4\u8272\u8499\u7248]
      \n F –> G[\u53cc\u8fb9\u6ee4\u6ce2\u78e8\u76ae]
      \n G –> H[HSV\u7f8e\u767d<br>V\u901a\u9053+15~30]
      \n H –> I[CLAHE\u5bf9\u6bd4\u5ea6\u589e\u5f3a]
      \n I –> J[\u8499\u7248\u878d\u5408<br>\u4fdd\u7559\u4e94\u5b98\u7ec6\u8282]
      \n J –> K[\u8f93\u51fa\u7f8e\u989c\u56fe\u50cf]<\/p>\n

      4.2 \u6838\u5fc3\u4ee3\u7801\u5b9e\u73b0<\/h4>\n

      import<\/span> cv2
      \nimport<\/span> numpy as<\/span> np
      \nfrom<\/span> dlib import<\/span> get_frontal_face_detector,<\/span> shape_predictor<\/p>\n

      class<\/span> OpenCVBeautifier<\/span>:<\/span>
      \n def<\/span> __init__<\/span>(<\/span>self)<\/span>:<\/span>
      \n # \u52a0\u8f7d\u4eba\u8138\u68c0\u6d4b\u4e0e\u7279\u5f81\u70b9\u6a21\u578b<\/span>
      \n self.<\/span>face_detector =<\/span> get_frontal_face_detector(<\/span>)<\/span>
      \n self.<\/span>landmark_predictor =<\/span> shape_predictor(<\/span>"shape_predictor_68_face_landmarks.dat"<\/span>)<\/span><\/p>\n

      def<\/span> detect_face_landmarks<\/span>(<\/span>self,<\/span> image)<\/span>:<\/span>
      \n """\u68c0\u6d4b\u4eba\u813868\u4e2a\u7279\u5f81\u70b9"""<\/span>
      \n gray =<\/span> cv2.<\/span>cvtColor(<\/span>image,<\/span> cv2.<\/span>COLOR_BGR2GRAY)<\/span>
      \n faces =<\/span> self.<\/span>face_detector(<\/span>gray)<\/span>
      \n landmarks_list =<\/span> [<\/span>]<\/span>
      \n for<\/span> face in<\/span> faces:<\/span>
      \n landmarks =<\/span> self.<\/span>landmark_predictor(<\/span>gray,<\/span> face)<\/span>
      \n landmarks_list.<\/span>append(<\/span>[<\/span>(<\/span>p.<\/span>x,<\/span> p.<\/span>y)<\/span> for<\/span> p in<\/span> landmarks.<\/span>parts(<\/span>)<\/span>]<\/span>)<\/span>
      \n return<\/span> landmarks_list<\/p>\n

      def<\/span> beautify<\/span>(<\/span>self,<\/span> image,<\/span> sigma_color=<\/span>35<\/span>,<\/span> sigma_space=<\/span>20<\/span>,<\/span> brightness=<\/span>20<\/span>)<\/span>:<\/span>
      \n """\u5b8c\u6574\u7f8e\u989c\u6d41\u7a0b"""<\/span>
      \n # 1. \u4eba\u8138\u4e0e\u7279\u5f81\u70b9\u68c0\u6d4b<\/span>
      \n landmarks_list =<\/span> self.<\/span>detect_face_landmarks(<\/span>image)<\/span>
      \n if<\/span> not<\/span> landmarks_list:<\/span>
      \n return<\/span> image<\/p>\n

      # 2. \u751f\u6210\u80a4\u8272\u8499\u7248<\/span>
      \n mask =<\/span> create_skin_mask(<\/span>image)<\/span><\/p>\n

      # 3. \u78e8\u76ae\u5904\u7406<\/span>
      \n smoothed =<\/span> bilateral_filter_skin_smoothing(<\/span>image,<\/span> sigma_color,<\/span> sigma_space)<\/span><\/p>\n

      # 4. \u7f8e\u767d（HSV\u7a7a\u95f4V\u901a\u9053\u8c03\u6574）<\/span>
      \n hsv =<\/span> cv2.<\/span>cvtColor(<\/span>smoothed,<\/span> cv2.<\/span>COLOR_BGR2HSV)<\/span>
      \n hsv[<\/span>:<\/span>,<\/span> :<\/span>,<\/span> 2<\/span>]<\/span> =<\/span> np.<\/span>clip(<\/span>hsv[<\/span>:<\/span>,<\/span> :<\/span>,<\/span> 2<\/span>]<\/span> +<\/span> brightness,<\/span> 0<\/span>,<\/span> 255<\/span>)<\/span>
      \n \u7f8e\u767d_image =<\/span> cv2.<\/span>cvtColor(<\/span>hsv,<\/span> cv2.<\/span>COLOR_HSV2BGR)<\/span><\/p>\n

      # 5. \u878d\u5408\u8499\u7248（\u4ec5\u4fdd\u7559\u76ae\u80a4\u533a\u57df\u7f8e\u989c\u6548\u679c）<\/span>
      \n result =<\/span> np.<\/span>where(<\/span>mask[<\/span>.<\/span>.<\/span>.<\/span>,<\/span> None<\/span>]<\/span> ==<\/span> 255<\/span>,<\/span> \u7f8e\u767d_image,<\/span> image)<\/span><\/p>\n

      return<\/span> result<\/p>\n

      # \u4f7f\u7528\u793a\u4f8b<\/span>
      \nif<\/span> __name__ ==<\/span> "__main__"<\/span>:<\/span>
      \n beautifier =<\/span> OpenCVBeautifier(<\/span>)<\/span>
      \n image =<\/span> cv2.<\/span>imread(<\/span>"input.jpg"<\/span>)<\/span>
      \n result =<\/span> beautifier.<\/span>beautify(<\/span>image)<\/span>
      \n cv2.<\/span>imwrite(<\/span>"beautified.jpg"<\/span>,<\/span> result)<\/span>
      \n cv2.<\/span>imshow(<\/span>"Original vs Beautified"<\/span>,<\/span> np.<\/span>hstack(<\/span>[<\/span>image,<\/span> result]<\/span>)<\/span>)<\/span>
      \n cv2.<\/span>waitKey(<\/span>0<\/span>)<\/span><\/p>\n

      \u4e94\u3001\u6027\u80fd\u4f18\u5316\u4e0e\u6700\u4f73\u5b9e\u8df5<\/h3>\n

      5.1 GPU\u52a0\u901f\u5173\u952e\u6280\u5de7<\/h4>\n
        \n
      • CUDA\u6a21\u5757\u8c03\u7528：\u4f7f\u7528cv2.cuda.bilateralFilter\u66ff\u4ee3CPU\u7248\u672c，\u5904\u74061080P\u56fe\u50cf\u8017\u65f6\u4ece80ms\u964d\u81f312ms<\/li>\n
      • \u5185\u5b58\u4f18\u5316：\u901a\u8fc7cv2.UMat\u5b9e\u73b0CPU\/GPU\u5185\u5b58\u81ea\u52a8\u5207\u6362，\u51cf\u5c11\u6570\u636e\u4f20\u8f93\u5f00\u9500<\/li>\n<\/ul>\n

        # GPU\u52a0\u901f\u793a\u4f8b<\/span>
        \ndef<\/span> gpu_bilateral_filter<\/span>(<\/span>image)<\/span>:<\/span>
        \n gpu_img =<\/span> cv2.<\/span>cuda_GpuMat(<\/span>)<\/span>
        \n gpu_img.<\/span>upload(<\/span>image)<\/span>
        \n gpu_result =<\/span> cv2.<\/span>cuda.<\/span>bilateralFilter(<\/span>gpu_img,<\/span> 0<\/span>,<\/span> 30<\/span>,<\/span> 15<\/span>)<\/span>
        \n return<\/span> gpu_result.<\/span>download(<\/span>)<\/span><\/p>\n

        5.2 \u53c2\u6570\u8c03\u4f18\u6307\u5357<\/h4>\n\n\u7f8e\u989c\u6548\u679c\u6838\u5fc3\u53c2\u6570\u63a8\u8350\u8303\u56f4<\/tr>\n\n\n\n\n\n
        \u78e8\u76ae\u5f3a\u5ea6<\/td>\nsigmaColor<\/td>\n30-50（\u8d8a\u9ad8\u8d8a\u6a21\u7cca）<\/td>\n<\/tr>\n
        \u8fb9\u7f18\u4fdd\u7559\u7a0b\u5ea6<\/td>\nsigmaSpace<\/td>\n15-30（\u8d8a\u4f4e\u8d8a\u9510\u5229）<\/td>\n<\/tr>\n
        \u7f8e\u767d\u7a0b\u5ea6<\/td>\nHSV-V\u901a\u9053\u589e\u91cf<\/td>\n15-30（\u907f\u514d\u8fc7\u66dd）<\/td>\n<\/tr>\n
        \u5b9e\u65f6\u6027\u6743\u8861<\/td>\n\u56fe\u50cf\u5206\u8fa8\u7387<\/td>\n720P（\u5e73\u8861\u6548\u679c\u4e0e\u901f\u5ea6）<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

        \u516d\u3001\u5f00\u6e90\u9879\u76ee\u4e0e\u6269\u5c55\u65b9\u5411<\/h3>\n

        6.1 \u63a8\u8350\u5b66\u4e60\u8d44\u6e90<\/h4>\n
          \n
        • OpenCV\u5b98\u65b9\u793a\u4f8b：G-API\u4eba\u8138\u7f8e\u989c\u6559\u7a0b<\/li>\n
        • \u5f00\u6e90\u9879\u76ee：\n
            \n
          • RealTimeNeuralStyleBeautify（\u5b9e\u65f6\u89c6\u9891\u7f8e\u989c）<\/li>\n
          • face_beauty（\u96c6\u6210\u5927\u773c\u3001\u7626\u8138\u7b97\u6cd5）<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n

            6.2 \u8fdb\u9636\u65b9\u5411<\/h4>\n
          • \u57fa\u4e8eGAN\u7684\u8d85\u5206\u8fa8\u7387：\u7ed3\u5408ESRGAN\u63d0\u5347\u7f8e\u989c\u540e\u7ec6\u8282<\/li>\n
          • 3D\u4eba\u8138\u5f62\u53d8：\u901a\u8fc7Dlib\u7279\u5f81\u70b9\u5b9e\u73b0\u7626\u8138\u3001\u5927\u773c\u7b49\u51e0\u4f55\u53d8\u6362<\/li>\n
          • \u79fb\u52a8\u7aef\u4f18\u5316：\u4f7f\u7528TensorRT\u5c06\u6a21\u578b\u63a8\u7406\u5ef6\u8fdf\u538b\u7f29\u81f320ms\u5185<\/li>\n

            \u7ed3\u8bed<\/h3>\n

            OpenCV\u63d0\u4f9b\u4e86\u4ece\u4f20\u7edf\u6ee4\u6ce2\u5230\u5e95\u5c42\u6df1\u5ea6\u5b66\u4e60\u96c6\u6210\u7684\u5168\u6808\u7f8e\u989c\u6280\u672f\u652f\u6301\u3002\u5f00\u53d1\u8005\u53ef\u6839\u636e\u9700\u6c42\u9009\u62e9\u201c\u8f7b\u91cf\u7ea7\u4f20\u7edf\u7b97\u6cd5\u201d（\u9002\u5408\u79fb\u52a8\u7aef\u5b9e\u65f6\u573a\u666f）\u6216\u201c\u6df1\u5ea6\u5b66\u4e60\u6df7\u5408\u6a21\u578b\u201d（\u9002\u5408\u9ad8\u7cbe\u5ea6\u573a\u666f），\u5e76\u901a\u8fc7GPU\u52a0\u901f\u4e0e\u53c2\u6570\u8c03\u4f18\u5e73\u8861\u6548\u679c\u4e0e\u6027\u80fd\u0964 \u672a\u6765，\u968f\u7740\u5b9e\u65f6\u98ce\u683c\u8fc1\u79fb\u4e0e\u795e\u7ecf\u6e32\u67d3\u6280\u672f\u7684\u53d1\u5c55，OpenCV\u7f8e\u989c\u7cfb\u7edf\u5c06\u5728\u865a\u62df\u8bd5\u5986\u3001AR\u76f4\u64ad\u7b49\u9886\u57df\u53d1\u6325\u66f4\u5927\u6f5c\u529b\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"

            \u6587\u7ae0\u6d4f\u89c8\u9605\u8bfb937\u6b21\uff0c\u70b9\u8d5e32\u6b21\uff0c\u6536\u85cf28\u6b21\u3002\u57fa\u4e8eOpenCV\u7684\u5b9e\u65f6\u7f8e\u989c\u6280\u672f\uff1a\u4ece\u4f20\u7edf\u7b97\u6cd5\u5230\u6df1\u5ea6\u5b66\u4e60\u878d\u5408\u5b9e\u73b0_opencv\u89c6\u9891\u7f8e\u989c\u5f00\u6e90<\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[5962,1460,81,5961,5957,86,5958,427,5959,5960],"topic":[],"class_list":["post-57167","post","type-post","status-publish","format-standard","hentry","category-server","tag-gpu","tag-opencv","tag-python","tag-5961","tag--opencv","tag-86","tag--pytorch","tag-427","tag-5959","tag-5960"],"yoast_head":"\n\u57fa\u4e8eOpenCV\u7684\u5b9e\u65f6\u7f8e\u989c\u6280\u672f\uff1a\u4ece\u4f20\u7edf\u7b97\u6cd5\u5230\u6df1\u5ea6\u5b66\u4e60\u878d\u5408\u5b9e\u73b0 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.wsisp.com\/helps\/57167.html\" \/>\n<meta property=\"og:locale\" content=\"zh_CN\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"\u57fa\u4e8eOpenCV\u7684\u5b9e\u65f6\u7f8e\u989c\u6280\u672f\uff1a\u4ece\u4f20\u7edf\u7b97\u6cd5\u5230\u6df1\u5ea6\u5b66\u4e60\u878d\u5408\u5b9e\u73b0 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"og:description\" content=\"\u6587\u7ae0\u6d4f\u89c8\u9605\u8bfb937\u6b21\uff0c\u70b9\u8d5e32\u6b21\uff0c\u6536\u85cf28\u6b21\u3002\u57fa\u4e8eOpenCV\u7684\u5b9e\u65f6\u7f8e\u989c\u6280\u672f\uff1a\u4ece\u4f20\u7edf\u7b97\u6cd5\u5230\u6df1\u5ea6\u5b66\u4e60\u878d\u5408\u5b9e\u73b0_opencv\u89c6\u9891\u7f8e\u989c\u5f00\u6e90\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.wsisp.com\/helps\/57167.html\" \/>\n<meta property=\"og:site_name\" content=\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"article:published_time\" content=\"2025-08-15T00:20:05+00:00\" \/>\n<meta name=\"author\" content=\"admin\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u4f5c\u8005\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin\" \/>\n\t<meta name=\"twitter:label2\" content=\"\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4\" \/>\n\t<meta name=\"twitter:data2\" content=\"4 \u5206\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/57167.html\",\"url\":\"https:\/\/www.wsisp.com\/helps\/57167.html\",\"name\":\"\u57fa\u4e8eOpenCV\u7684\u5b9e\u65f6\u7f8e\u989c\u6280\u672f\uff1a\u4ece\u4f20\u7edf\u7b97\u6cd5\u5230\u6df1\u5ea6\u5b66\u4e60\u878d\u5408\u5b9e\u73b0 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"isPartOf\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\"},\"datePublished\":\"2025-08-15T00:20:05+00:00\",\"dateModified\":\"2025-08-15T00:20:05+00:00\",\"author\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\"},\"breadcrumb\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/57167.html#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.wsisp.com\/helps\/57167.html\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/57167.html#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"\u9996\u9875\",\"item\":\"https:\/\/www.wsisp.com\/helps\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"\u57fa\u4e8eOpenCV\u7684\u5b9e\u65f6\u7f8e\u989c\u6280\u672f\uff1a\u4ece\u4f20\u7edf\u7b97\u6cd5\u5230\u6df1\u5ea6\u5b66\u4e60\u878d\u5408\u5b9e\u73b0\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\",\"url\":\"https:\/\/www.wsisp.com\/helps\/\",\"name\":\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"description\":\"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"zh-Hans\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\",\"name\":\"admin\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"contentUrl\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"caption\":\"admin\"},\"sameAs\":[\"http:\/\/wp.wsisp.com\"],\"url\":\"https:\/\/www.wsisp.com\/helps\/author\/admin\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"\u57fa\u4e8eOpenCV\u7684\u5b9e\u65f6\u7f8e\u989c\u6280\u672f\uff1a\u4ece\u4f20\u7edf\u7b97\u6cd5\u5230\u6df1\u5ea6\u5b66\u4e60\u878d\u5408\u5b9e\u73b0 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.wsisp.com\/helps\/57167.html","og_locale":"zh_CN","og_type":"article","og_title":"\u57fa\u4e8eOpenCV\u7684\u5b9e\u65f6\u7f8e\u989c\u6280\u672f\uff1a\u4ece\u4f20\u7edf\u7b97\u6cd5\u5230\u6df1\u5ea6\u5b66\u4e60\u878d\u5408\u5b9e\u73b0 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","og_description":"\u6587\u7ae0\u6d4f\u89c8\u9605\u8bfb937\u6b21\uff0c\u70b9\u8d5e32\u6b21\uff0c\u6536\u85cf28\u6b21\u3002\u57fa\u4e8eOpenCV\u7684\u5b9e\u65f6\u7f8e\u989c\u6280\u672f\uff1a\u4ece\u4f20\u7edf\u7b97\u6cd5\u5230\u6df1\u5ea6\u5b66\u4e60\u878d\u5408\u5b9e\u73b0_opencv\u89c6\u9891\u7f8e\u989c\u5f00\u6e90","og_url":"https:\/\/www.wsisp.com\/helps\/57167.html","og_site_name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","article_published_time":"2025-08-15T00:20:05+00:00","author":"admin","twitter_card":"summary_large_image","twitter_misc":{"\u4f5c\u8005":"admin","\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4":"4 \u5206"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/www.wsisp.com\/helps\/57167.html","url":"https:\/\/www.wsisp.com\/helps\/57167.html","name":"\u57fa\u4e8eOpenCV\u7684\u5b9e\u65f6\u7f8e\u989c\u6280\u672f\uff1a\u4ece\u4f20\u7edf\u7b97\u6cd5\u5230\u6df1\u5ea6\u5b66\u4e60\u878d\u5408\u5b9e\u73b0 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","isPartOf":{"@id":"https:\/\/www.wsisp.com\/helps\/#website"},"datePublished":"2025-08-15T00:20:05+00:00","dateModified":"2025-08-15T00:20:05+00:00","author":{"@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41"},"breadcrumb":{"@id":"https:\/\/www.wsisp.com\/helps\/57167.html#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.wsisp.com\/helps\/57167.html"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/www.wsisp.com\/helps\/57167.html#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"\u9996\u9875","item":"https:\/\/www.wsisp.com\/helps"},{"@type":"ListItem","position":2,"name":"\u57fa\u4e8eOpenCV\u7684\u5b9e\u65f6\u7f8e\u989c\u6280\u672f\uff1a\u4ece\u4f20\u7edf\u7b97\u6cd5\u5230\u6df1\u5ea6\u5b66\u4e60\u878d\u5408\u5b9e\u73b0"}]},{"@type":"WebSite","@id":"https:\/\/www.wsisp.com\/helps\/#website","url":"https:\/\/www.wsisp.com\/helps\/","name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","description":"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"zh-Hans"},{"@type":"Person","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41","name":"admin","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/","url":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","contentUrl":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","caption":"admin"},"sameAs":["http:\/\/wp.wsisp.com"],"url":"https:\/\/www.wsisp.com\/helps\/author\/admin"}]}},"_links":{"self":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/57167","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/comments?post=57167"}],"version-history":[{"count":0,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/57167\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media?parent=57167"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/categories?post=57167"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/tags?post=57167"},{"taxonomy":"topic","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/topic?post=57167"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}