{"id":1014790,"date":"2024-12-27T11:59:28","date_gmt":"2024-12-27T03:59:28","guid":{"rendered":"https:\/\/docs.pingcode.com\/ask\/ask-ask\/1014790.html"},"modified":"2024-12-27T11:59:30","modified_gmt":"2024-12-27T03:59:30","slug":"%e7%94%a8python%e5%a6%82%e4%bd%95%e6%8a%a0%e5%9b%be","status":"publish","type":"post","link":"https:\/\/docs.pingcode.com\/ask\/1014790.html","title":{"rendered":"\u7528python\u5982\u4f55\u62a0\u56fe"},"content":{"rendered":"

\"\u7528python\u5982\u4f55\u62a0\u56fe\"<\/p>\n

\u5f00\u5934\u6bb5\u843d:
\u4f7f\u7528Python\u8fdb\u884c\u62a0\u56fe\u4e3b\u8981\u4f9d\u8d56\u4e8e\u56fe\u50cf\u5904\u7406\u5e93\uff0c\u5982OpenCV\u3001Pillow\u3001Scikit-Image<\/strong>\u7b49\u3002OpenCV\u63d0\u4f9b\u4e86\u5f3a\u5927\u7684\u56fe\u50cf\u5904\u7406\u529f\u80fd\uff0c\u53ef\u4ee5\u901a\u8fc7\u8fb9\u7f18\u68c0\u6d4b\u3001\u989c\u8272\u5206\u5272\u7b49\u65b9\u6cd5\u5b9e\u73b0\u62a0\u56fe\uff1bPillow\u5219\u63d0\u4f9b\u4e86\u7b80\u5355\u7684\u56fe\u50cf\u52a0\u8f7d\u548c\u5904\u7406\u529f\u80fd\uff0c\u9002\u7528\u4e8e\u8f7b\u91cf\u7ea7\u7684\u62a0\u56fe\u64cd\u4f5c\uff1b\u800cScikit-Image\u5219\u652f\u6301\u590d\u6742\u7684\u56fe\u50cf\u5206\u5272\u7b97\u6cd5\uff0c\u5982\u56fe\u5272\u7b97\u6cd5\u3002\u8fd9\u4e9b\u5e93\u5404\u6709\u4f18\u52a3\uff0c\u9002\u5408\u4e0d\u540c\u7684\u5e94\u7528\u573a\u666f\u3002\u63a5\u4e0b\u6765\uff0c\u6211\u5c06\u8be6\u7ec6\u4ecb\u7ecd\u5982\u4f55\u4f7f\u7528OpenCV\u8fdb\u884c\u62a0\u56fe\u64cd\u4f5c\u3002<\/p>\n<\/p>\n

\u4e00\u3001OPENCV\u8fdb\u884c\u62a0\u56fe<\/p>\n<\/p>\n

OpenCV\u662f\u4e00\u4e2a\u5f00\u6e90\u7684\u8ba1\u7b97\u673a\u89c6\u89c9\u5e93\uff0c\u5177\u6709\u4e30\u5bcc\u7684\u56fe\u50cf\u5904\u7406\u529f\u80fd\u3002\u8981\u4f7f\u7528OpenCV\u8fdb\u884c\u62a0\u56fe\uff0c\u6211\u4eec\u9996\u5148\u9700\u8981\u8fdb\u884c\u56fe\u50cf\u8bfb\u53d6\uff0c\u7136\u540e\u8fdb\u884c\u9884\u5904\u7406\uff08\u5982\u8f6c\u4e3a\u7070\u5ea6\u56fe\u3001\u6a21\u7cca\u5904\u7406\u7b49\uff09\uff0c\u6700\u540e\u8fdb\u884c\u56fe\u50cf\u5206\u5272\u3002<\/p>\n<\/p>\n

    \n
  1. \u56fe\u50cf\u8bfb\u53d6\u4e0e\u9884\u5904\u7406<\/strong><\/li>\n<\/ol>\n

    \u9996\u5148\uff0c\u6211\u4eec\u9700\u8981\u4f7f\u7528OpenCV\u8bfb\u53d6\u56fe\u50cf\u5e76\u8fdb\u884c\u4e00\u4e9b\u57fa\u672c\u7684\u9884\u5904\u7406\u3002\u5e38\u7528\u7684\u9884\u5904\u7406\u65b9\u6cd5\u5305\u62ec\u7070\u5ea6\u8f6c\u6362\u3001\u6a21\u7cca\u5904\u7406\u7b49\uff0c\u4ee5\u4fbf\u4e8e\u540e\u7eed\u7684\u56fe\u50cf\u5904\u7406\u3002<\/p>\n<\/p>\n

    import cv2<\/p>\n
    \u8bfb\u53d6\u56fe\u50cf<\/strong><\/h2>\n
    image = cv2.imread('image.jpg')<\/p>\n
    \u8f6c\u6362\u4e3a\u7070\u5ea6\u56fe<\/strong><\/h2>\n
    gray_image = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)<\/p>\n
    \u5e94\u7528\u9ad8\u65af\u6a21\u7cca<\/strong><\/h2>\n
    blurred_image = cv2.GaussianBlur(gray_image, (5, 5), 0)<\/p>\n
    <\/code><\/pre>\n<\/p>\n
      \n
    1. \u8fb9\u7f18\u68c0\u6d4b\u4e0e\u8f6e\u5ed3\u67e5\u627e<\/strong><\/li>\n<\/ol>\n
      \u901a\u8fc7\u8fb9\u7f18\u68c0\u6d4b\uff0c\u6211\u4eec\u53ef\u4ee5\u627e\u5230\u56fe\u50cf\u4e2d\u7684\u7269\u4f53\u8fb9\u754c\u3002\u5e38\u7528\u7684\u8fb9\u7f18\u68c0\u6d4b\u7b97\u6cd5\u5305\u62ecCanny\u8fb9\u7f18\u68c0\u6d4b\u7b49\u3002\u968f\u540e\uff0c\u6211\u4eec\u53ef\u4ee5\u4f7f\u7528\u8f6e\u5ed3\u67e5\u627e\u65b9\u6cd5\u63d0\u53d6\u51fa\u7269\u4f53\u7684\u8f6e\u5ed3\u3002<\/p>\n<\/p>\n
      # \u4f7f\u7528Canny\u8fb9\u7f18\u68c0\u6d4b<\/p>\n
      edges = cv2.Canny(blurred_image, 50, 150)<\/p>\n
      \u67e5\u627e\u8f6e\u5ed3<\/strong><\/h2>\n
      contours, _ = cv2.findContours(edges, cv2.RETR_EXTERNAL, cv2.CHAI<\/a>N_APPROX_SIMPLE)<\/p>\n
      <\/code><\/pre>\n<\/p>\n
        \n
      1. \u56fe\u50cf\u5206\u5272\u4e0e\u62a0\u56fe<\/strong><\/li>\n<\/ol>\n
        \u5728\u83b7\u5f97\u7269\u4f53\u7684\u8f6e\u5ed3\u540e\uff0c\u6211\u4eec\u53ef\u4ee5\u901a\u8fc7\u8499\u7248\u64cd\u4f5c\u5b9e\u73b0\u56fe\u50cf\u5206\u5272\uff0c\u4ece\u800c\u5b9e\u73b0\u62a0\u56fe\u3002<\/p>\n<\/p>\n
        # \u521b\u5efa\u4e00\u4e2a\u9ed1\u8272\u80cc\u666f\u7684\u63a9\u819c<\/p>\n
        mask = np.zeros_like(image)<\/p>\n
        \u7ed8\u5236\u8f6e\u5ed3<\/strong><\/h2>\n
        cv2.drawContours(mask, contours, -1, (255, 255, 255), thickness=cv2.FILLED)<\/p>\n
        \u5e94\u7528\u63a9\u819c<\/strong><\/h2>\n
        segmented_image = cv2.bitwise_and(image, mask)<\/p>\n
        <\/code><\/pre>\n<\/p>\n
        \u4e8c\u3001PILLOW\u8fdb\u884c\u62a0\u56fe<\/p>\n<\/p>\n
        Pillow\u662fPython Imaging Library\u7684\u5206\u652f\uff0c\u662f\u4e00\u4e2a\u53cb\u597d\u7684\u56fe\u50cf\u5904\u7406\u5e93\u3002\u5b83\u9002\u5408\u8fdb\u884c\u7b80\u5355\u7684\u56fe\u50cf\u5904\u7406\u64cd\u4f5c\uff0c\u5982\u62a0\u56fe\u3002<\/p>\n<\/p>\n
          \n
        1. \u56fe\u50cf\u52a0\u8f7d\u4e0e\u7b80\u5355\u5904\u7406<\/strong><\/li>\n<\/ol>\n
          \u4f7f\u7528Pillow\u52a0\u8f7d\u56fe\u50cf\uff0c\u5e76\u8fdb\u884c\u57fa\u672c\u7684\u56fe\u50cf\u5904\u7406\u64cd\u4f5c\uff0c\u5982\u8c03\u6574\u5927\u5c0f\u3001\u65cb\u8f6c\u7b49\u3002<\/p>\n<\/p>\n
          from PIL import Image<\/p>\n
          \u6253\u5f00\u56fe\u50cf<\/strong><\/h2>\n
          image = Image.open('image.jpg')<\/p>\n
          \u8c03\u6574\u56fe\u50cf\u5927\u5c0f<\/strong><\/h2>\n
          image = image.resize((300, 300))<\/p>\n
          \u65cb\u8f6c\u56fe\u50cf<\/strong><\/h2>\n
          image = image.rotate(45)<\/p>\n
          <\/code><\/pre>\n<\/p>\n
            \n
          1. \u4f7f\u7528\u989c\u8272\u5206\u5272\u8fdb\u884c\u62a0\u56fe<\/strong><\/li>\n<\/ol>\n
            Pillow\u652f\u6301\u7b80\u5355\u7684\u989c\u8272\u5206\u5272\u6280\u672f\uff0c\u53ef\u4ee5\u7528\u4e8e\u62a0\u56fe\u3002\u901a\u8fc7\u8bbe\u7f6e\u989c\u8272\u9608\u503c\uff0c\u6211\u4eec\u53ef\u4ee5\u5206\u5272\u51fa\u7279\u5b9a\u989c\u8272\u7684\u533a\u57df\u3002<\/p>\n<\/p>\n
            from PIL import ImageChops<\/p>\n
            \u5206\u5272\u56fe\u50cf<\/strong><\/h2>\n
            def split_image(image, threshold=100):<\/p>\n
                # \u4f7f\u7528\u7070\u5ea6<\/p>\n
                grayscale = image.convert("L")<\/p>\n
                # \u521b\u5efa\u63a9\u819c<\/p>\n
                mask = grayscale.point(lambda x: 255 if x < threshold else 0)<\/p>\n
                # \u5e94\u7528\u63a9\u819c<\/p>\n
                return ImageChops.multiply(image, Image.merge("RGB", [mask, mask, mask]))<\/p>\n
            segmented_image = split_image(image)<\/p>\n
            <\/code><\/pre>\n<\/p>\n
            \u4e09\u3001SCIKIT-IMAGE\u8fdb\u884c\u62a0\u56fe<\/p>\n<\/p>\n
            Scikit-Image\u662f\u4e00\u4e2a\u7528\u4e8e\u56fe\u50cf\u5904\u7406\u7684Python\u5e93\uff0c\u652f\u6301\u590d\u6742\u7684\u56fe\u50cf\u5206\u5272\u7b97\u6cd5\uff0c\u5982\u56fe\u5272\u3001\u533a\u57df\u751f\u957f\u7b49\u3002<\/p>\n<\/p>\n
              \n
            1. \u56fe\u50cf\u52a0\u8f7d\u4e0e\u9884\u5904\u7406<\/strong><\/li>\n<\/ol>\n
              \u9996\u5148\uff0c\u4f7f\u7528Scikit-Image\u52a0\u8f7d\u56fe\u50cf\u5e76\u8fdb\u884c\u9884\u5904\u7406\u3002\u53ef\u4ee5\u901a\u8fc7\u8c03\u6574\u56fe\u50cf\u7684\u989c\u8272\u7a7a\u95f4\u3001\u5e94\u7528\u6ee4\u6ce2\u5668\u7b49\u6765\u7b80\u5316\u56fe\u50cf\u5904\u7406\u3002<\/p>\n<\/p>\n
              from skimage import io, color, filters<\/p>\n
              \u8bfb\u53d6\u56fe\u50cf<\/strong><\/h2>\n
              image = io.imread('image.jpg')<\/p>\n
              \u8f6c\u6362\u4e3a\u7070\u5ea6\u56fe<\/strong><\/h2>\n
              gray_image = color.rgb2gray(image)<\/p>\n
              \u5e94\u7528\u4e2d\u503c\u6ee4\u6ce2<\/strong><\/h2>\n
              filtered_image = filters.median(gray_image)<\/p>\n
              <\/code><\/pre>\n<\/p>\n
                \n
              1. \u56fe\u50cf\u5206\u5272<\/strong><\/li>\n<\/ol>\n
                Scikit-Image\u63d0\u4f9b\u4e86\u591a\u79cd\u56fe\u50cf\u5206\u5272\u7b97\u6cd5\uff0c\u53ef\u4ee5\u6839\u636e\u56fe\u50cf\u7279\u70b9\u9009\u62e9\u5408\u9002\u7684\u7b97\u6cd5\u8fdb\u884c\u62a0\u56fe\u3002<\/p>\n<\/p>\n
                from skimage import segmentation, measure<\/p>\n
                \u4f7f\u7528\u5206\u6c34\u5cad\u7b97\u6cd5\u8fdb\u884c\u56fe\u50cf\u5206\u5272<\/strong><\/h2>\n
                markers = filters.sobel(filtered_image)<\/p>\n
                segmented_image = segmentation.watershed(filtered_image, markers)<\/p>\n
                \u67e5\u627e\u8f6e\u5ed3<\/strong><\/h2>\n
                contours = measure.find_contours(segmented_image, 0.8)<\/p>\n
                <\/code><\/pre>\n<\/p>\n
                  \n
                1. \u5e94\u7528\u8499\u7248\u8fdb\u884c\u62a0\u56fe<\/strong><\/li>\n<\/ol>\n
                  \u901a\u8fc7\u627e\u5230\u7684\u8f6e\u5ed3\uff0c\u53ef\u4ee5\u521b\u5efa\u8499\u7248\u6765\u5b9e\u73b0\u62a0\u56fe\u64cd\u4f5c\u3002<\/p>\n<\/p>\n
                  import numpy as np<\/p>\n
                  \u521b\u5efa\u8499\u7248<\/strong><\/h2>\n
                  mask = np.zeros_like(image, dtype=bool)<\/p>\n
                  \u5e94\u7528\u8f6e\u5ed3<\/strong><\/h2>\n
                  for contour in contours:<\/p>\n
                      for point in contour:<\/p>\n
                          mask[int(point[0]), int(point[1])] = True<\/p>\n
                  \u5e94\u7528\u8499\u7248<\/strong><\/h2>\n
                  segmented_image = np.zeros_like(image)<\/p>\n
                  segmented_image[mask] = image[mask]<\/p>\n
                  <\/code><\/pre>\n<\/p>\n
                  \u56db\u3001\u603b\u7ed3<\/p>\n<\/p>\n
                  Python\u63d0\u4f9b\u4e86\u591a\u79cd\u56fe\u50cf\u5904\u7406\u5e93\u6765\u8fdb\u884c\u62a0\u56fe\u64cd\u4f5c\uff0c\u9009\u62e9\u5408\u9002\u7684\u5e93\u53ef\u4ee5\u6839\u636e\u5177\u4f53\u9700\u6c42\u548c\u56fe\u50cf\u7279\u70b9\u8fdb\u884c\u3002OpenCV\u9002\u5408\u590d\u6742\u7684\u56fe\u50cf\u5904\u7406\u4efb\u52a1\uff0cPillow\u9002\u5408\u7b80\u5355\u7684\u56fe\u50cf\u64cd\u4f5c\uff0cScikit-Image\u5219\u63d0\u4f9b\u4e86\u4e30\u5bcc\u7684\u56fe\u50cf\u5206\u5272\u7b97\u6cd5<\/strong>\u3002\u5728\u5b9e\u9645\u5e94\u7528\u4e2d\uff0c\u53ef\u4ee5\u7ed3\u5408\u591a\u79cd\u65b9\u6cd5\u6765\u5b9e\u73b0\u66f4\u7cbe\u786e\u7684\u62a0\u56fe\u6548\u679c\u3002<\/p>\n<\/p>\n
                  \u76f8\u5173\u95ee\u7b54FAQs\uff1a<\/strong><\/h2>\n
                   \u5982\u4f55\u4f7f\u7528Python\u8fdb\u884c\u62a0\u56fe\uff1f<\/strong>
                  \u5728Python\u4e2d\uff0c\u53ef\u4ee5\u4f7f\u7528\u5404\u79cd\u56fe\u50cf\u5904\u7406\u5e93\u6765\u5b9e\u73b0\u62a0\u56fe\uff0c\u5982OpenCV\u3001Pillow\u548cscikit-image\u7b49\u3002\u901a\u5e38\uff0c\u62a0\u56fe\u7684\u57fa\u672c\u6b65\u9aa4\u5305\u62ec\u8bfb\u53d6\u56fe\u50cf\u3001\u9009\u62e9\u611f\u5174\u8da3\u7684\u533a\u57df\uff08ROI\uff09\u3001\u5e94\u7528\u63a9\u819c\uff0c\u5e76\u6700\u7ec8\u4fdd\u5b58\u6216\u663e\u793a\u7ed3\u679c\u3002\u5177\u4f53\u5b9e\u73b0\u53ef\u4ee5\u901a\u8fc7\u8c03\u6574\u9608\u503c\u3001\u4f7f\u7528\u8fb9\u7f18\u68c0\u6d4b\u6216\u5206\u5272\u7b97\u6cd5\u6765\u83b7\u53d6\u7cbe\u786e\u7684\u62a0\u56fe\u6548\u679c\u3002<\/p>\n
                  \u54ea\u4e9bPython\u5e93\u9002\u5408\u7528\u4e8e\u62a0\u56fe\uff1f<\/strong>
                  \u5e38\u7528\u7684Python\u5e93\u5305\u62ecOpenCV\u3001Pillow\u3001scikit-image\u548cnumpy\u7b49\u3002OpenCV\u63d0\u4f9b\u4e86\u4e30\u5bcc\u7684\u56fe\u50cf\u5904\u7406\u529f\u80fd\uff0c\u9002\u7528\u4e8e\u590d\u6742\u7684\u62a0\u56fe\u9700\u6c42\u3002Pillow\u5219\u9002\u5408\u7b80\u5355\u7684\u56fe\u50cf\u64cd\u4f5c\u3002scikit-image\u4e13\u6ce8\u4e8e\u56fe\u50cf\u5206\u6790\u548c\u5904\u7406\uff0c\u529f\u80fd\u5f3a\u5927\uff0c\u53ef\u4ee5\u8fdb\u884c\u9ad8\u7ea7\u7684\u56fe\u50cf\u5206\u5272\u3002<\/p>\n
                  \u62a0\u56fe\u65f6\u5982\u4f55\u9009\u62e9\u80cc\u666f\u548c\u524d\u666f\uff1f<\/strong>
                  \u5728\u62a0\u56fe\u8fc7\u7a0b\u4e2d\uff0c\u9009\u62e9\u80cc\u666f\u548c\u524d\u666f\u901a\u5e38\u4f9d\u8d56\u4e8e\u5bf9\u56fe\u50cf\u5185\u5bb9\u7684\u7406\u89e3\u3002\u53ef\u4ee5\u4f7f\u7528\u989c\u8272\u9608\u503c\u5206\u5272\u3001\u8fb9\u7f18\u68c0\u6d4b\u6216\u6df1\u5ea6\u5b66\u4e60\u6a21\u578b\u6765\u5b9e\u73b0\u3002\u57fa\u4e8e\u989c\u8272\u7684\u9608\u503c\u5206\u5272\u53ef\u4ee5\u5e2e\u52a9\u5feb\u901f\u533a\u5206\u80cc\u666f\u4e0e\u524d\u666f\uff0c\u800c\u4f7f\u7528\u6df1\u5ea6\u5b66\u4e60\u6a21\u578b\uff08\u5982U-Net\u6216Mask R-CNN\uff09\u53ef\u4ee5\u83b7\u5f97\u66f4\u9ad8\u7684\u51c6\u786e\u5ea6\uff0c\u5c24\u5176\u662f\u5728\u590d\u6742\u573a\u666f\u4e0b\u3002 <\/p>\n
                  \u62a0\u56fe\u540e\u5982\u4f55\u4fdd\u5b58\u5904\u7406\u540e\u7684\u56fe\u50cf\uff1f<\/strong>
                  \u5904\u7406\u540e\u7684\u56fe\u50cf\u53ef\u4ee5\u4f7f\u7528Pillow\u5e93\u7684save()<\/code>\u65b9\u6cd5\u6216OpenCV\u7684imwrite()<\/code>\u51fd\u6570\u6765\u4fdd\u5b58\u3002\u786e\u4fdd\u5728\u4fdd\u5b58\u4e4b\u524d\u9009\u62e9\u5408\u9002\u7684\u56fe\u50cf\u683c\u5f0f\uff0c\u5982PNG\u6216JPEG\uff0c\u4ee5\u4fdd\u7559\u6240\u9700\u7684\u8d28\u91cf\u548c\u900f\u660e\u5ea6\u3002\u5982\u679c\u4f7f\u7528Pillow\uff0c\u53ef\u4ee5\u4f7f\u7528image.save("output.png")<\/code>\u8fdb\u884c\u4fdd\u5b58\uff1b\u800c\u5728OpenCV\u4e2d\uff0c\u4f7f\u7528cv2.imwrite("output.jpg", image)<\/code>\u5b8c\u6210\u4fdd\u5b58\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"\u5f00\u5934\u6bb5\u843d:\u4f7f\u7528Python\u8fdb\u884c\u62a0\u56fe\u4e3b\u8981\u4f9d\u8d56\u4e8e\u56fe\u50cf\u5904\u7406\u5e93\uff0c\u5982OpenCV\u3001Pillow\u3001Scikit-Image […]","protected":false},"author":3,"featured_media":1014800,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[37],"tags":[],"acf":[],"_links":{"self":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/1014790"}],"collection":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/comments?post=1014790"}],"version-history":[{"count":"1","href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/1014790\/revisions"}],"predecessor-version":[{"id":1014803,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/1014790\/revisions\/1014803"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/media\/1014800"}],"wp:attachment":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/media?parent=1014790"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/categories?post=1014790"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/tags?post=1014790"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}