优化图片分类和图片方向判断

util/image_util.py

@@ -1,6 +1,8 @@
+import logging
 import math
 
 import cv2
+import fitz
 import numpy
 from PIL import Image
 
@@ -27,7 +29,7 @@ def capture(image, rectangle):
 def split(img_path, ratio=1.414, overlap=0.05, x_compensation=3):
     """
     分割图片
-    :param img_path:图片路径
+    :param img_path: 图片路径
     :param ratio: 分割后的比例
     :param overlap: 图片之间的覆盖比例
     :param x_compensation: 横向补偿倍率
@@ -251,3 +253,33 @@ def is_photo(img_path):
         if any(tag in exif for tag in (271, 272)):
             return True
     return False
+
+
+def pdf_to_imgs(pdf_path, dpi=150):
+    pdf_file = None
+    # 定义缩放系数（DPI）
+    default_dpi = 72
+    zoom = dpi / default_dpi
+    try:
+        # 打开PDF文件
+        pdf_file = fitz.open(pdf_path)
+
+        pdf_imgs = []
+        for page in pdf_file:
+            # 设置矩阵变换参数
+            mat = fitz.Matrix(zoom, zoom)
+            # 渲染页面
+            pix = page.get_pixmap(matrix=mat)
+            # 将渲染结果转换为OpenCV兼容的格式
+            img = numpy.frombuffer(pix.samples, dtype=numpy.uint8).reshape((pix.height, pix.width, -1))
+            img = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)
+            pdf_imgs.append([img, page.get_text()])
+
+        return pdf_imgs
+
+    except Exception as ex:
+        logging.getLogger('error').error('pdf转图片失败！', exc_info=ex)
+        return None
+    finally:
+        if pdf_file:
+            pdf_file.close()