更新1.16.0版本

Dockerfile

@@ -15,6 +15,7 @@ ENV PYTHONUNBUFFERED=1 \
 COPY requirements.txt /app/requirements.txt
 COPY packages /app/packages
 RUN ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && echo '$TZ' > /etc/timezone \
+    && python3 -m pip install --upgrade pip \
     && pip install --no-cache-dir -r requirements.txt \
     && pip uninstall -y onnxruntime onnxruntime-gpu \
     && pip install onnxruntime-gpu==1.18.0 --extra-index-url https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/onnxruntime-cuda-12/pypi/simple/

README.md

@@ -129,4 +129,5 @@ bash update.sh
 22. 版本号：1.15.0
     1. 新增图片清晰度测试
 23. 版本号：1.16.0
-    1. 更新paddle框架至3.0
+    1. 优化结算单号规则
+    2. 新增判断截图方法

docker-compose.yml

@@ -1,6 +1,6 @@
 x-env:
   &template
-  image: fcb_photo_review:1.15.4
+  image: fcb_photo_review:1.16.0
   restart: always
 
 x-review:

photo_review/auto_photo_review.py

@@ -262,18 +262,18 @@ def information_extraction(ie, phrecs, identity):
         session.add_all(zx_ie_results)
         session.commit()
 
-        # 添加清晰度测试
-        if better_image is None:
-            # 替换后图片默认清晰
-            clarity_result = image_util.parse_clarity(image)
-            unsharp_flag = 0 if (clarity_result[0] == 0 and clarity_result[1] >= 0.8) else 1
-            update_clarity = (update(ZxPhrec).where(ZxPhrec.pk_phrec == phrec.pk_phrec).values(
-                cfjaddress2=json.dumps(clarity_result),
-                unsharp_flag=unsharp_flag,
-            ))
-            session.execute(update_clarity)
-            session.commit()
-        session.close()
+        # # 添加清晰度测试
+        # if better_image is None:
+        #     # 替换后图片默认清晰
+        #     clarity_result = image_util.parse_clarity(image)
+        #     unsharp_flag = 0 if (clarity_result[0] == 0 and clarity_result[1] >= 0.8) else 1
+        #     update_clarity = (update(ZxPhrec).where(ZxPhrec.pk_phrec == phrec.pk_phrec).values(
+        #         cfjaddress2=json.dumps(clarity_result),
+        #         unsharp_flag=unsharp_flag,
+        #     ))
+        #     session.execute(update_clarity)
+        #     session.commit()
+        # session.close()
 
     result['ocr_text'] = ocr_text
     return result
@@ -320,7 +320,7 @@ def save_or_update_ie(table, pk_phhd, data):
     if db_data:
         # 更新
         db_data.update_time = now
-        db_data.creator = HOSTNAME
+        db_data.updater = HOSTNAME
         for k, v in data.items():
             setattr(db_data, k, v)
     else:
@@ -421,6 +421,10 @@ def settlement_task(pk_phhd, settlement_list, identity):
                                      get_best_value_in_keys(settlement_list_ie_result, MEDICAL_EXPENSES))
     settlement_data["medical_expenses_str"] = handle_original_data(parse_money_result[0])
     settlement_data["medical_expenses"] = parse_money_result[1]
+
+    if not settlement_data["settlement_id"]:
+        # 如果没有结算单号就填住院号
+        settlement_data["settlement_id"] = settlement_data["admission_id"]
     save_or_update_ie(ZxIeSettlement, pk_phhd, settlement_data)
 
 

util/image_util.py

@@ -268,3 +268,61 @@ def parse_clarity(image):
     except Exception as e:
         logging.error("获取图片清晰度失败", exc_info=e)
     return clarity_result
+
+
+def is_photo_by_exif(exif_tags):
+    """分析EXIF数据判断是否为照片"""
+    # 照片通常包含的EXIF标签
+    photo_tags = [
+        'FNumber',  # 光圈
+        'ExposureTime',  # 曝光时间
+        'ISOSpeedRatings',  # ISO
+        'FocalLength',  # 焦距
+        'LensModel',  # 镜头型号
+        'GPSLatitude'  # GPS位置信息
+    ]
+
+    # 统计照片相关的EXIF标签数量
+    photo_tag_count = 0
+    if exif_tags:
+        for tag in photo_tags:
+            if tag in exif_tags:
+                photo_tag_count += 1
+                # 如果有2个以上照片相关的EXIF标签，倾向于是照片
+                if photo_tag_count >= 2:
+                    return True
+    # 不确定是照片返回False
+    return False
+
+
+def is_screenshot_by_image_features(image):
+    """分析图像特征判断是否为截图"""
+    # 定义边缘像素标准差阈值，小于此阈值则认为图片是截图
+    edge_std_threshold = 20.0
+    try:
+        # 检查边缘像素的一致性（截图边缘通常更整齐）
+        edge_pixels = []
+        # 取图像边缘10像素
+        edge_pixels.extend(image[:10, :].flatten())  # 顶部边缘
+        edge_pixels.extend(image[-10:, :].flatten())  # 底部边缘
+        edge_pixels.extend(image[:, :10].flatten())  # 左侧边缘
+        edge_pixels.extend(image[:, -10:].flatten())  # 右侧边缘
+
+        # 计算边缘像素的标准差（值越小说明越一致）
+        edge_std = numpy.std(edge_pixels)
+        logging.info(f"边缘像素标准差: {edge_std}")
+        return edge_std < edge_std_threshold
+    except Exception as e:
+        logging.error("图像特征分析失败", exc_info=e)
+        return False
+
+
+def is_screenshot(image, exif_tags):
+    """综合判断是否是截图"""
+    # 先检查EXIF数据
+    result_of_exif = is_photo_by_exif(exif_tags)
+    # 如果有明显的照片EXIF信息，直接判断为照片
+    if result_of_exif:
+        return False
+    # 分析图像特征
+    return is_screenshot_by_image_features(image)