Loading... 通过图片相似度批量删除重复图片的代码。 ## 首先安装必要的库 ``` pip install opencv-python numpy keras tensorflow ``` ## 运行脚本代码 ``` import os import cv2 import numpy as np from keras.applications.resnet50 import ResNet50, preprocess_input from pathlib import Path # 导入pathlib库 def extract_image_features(image_path): try: image = cv2.imread(image_path) # 读取图片 if image is None: print(f"无法读取图片: {image_path}") return None image = cv2.resize(image, (256, 256)) # 缩放图片到统一尺寸 image = image[16:240, 16:240] # 裁剪中间区域(224x224) image = np.expand_dims(image, axis=0) # 扩展维度以匹配模型输入要求 image = preprocess_input(image) # 预处理图片 features = model.predict(image) # 提取特征向量 features /= np.linalg.norm(features) # 归一化特征向量 return features.flatten() # 平铺特征向量 except Exception as e: print(f"处理图片时出错: {e}") return None def delete_duplicate_images(): current_dir = Path(os.getcwd()) # 使用pathlib获取当前目录路径 files = [f for f in current_dir.iterdir() if f.is_file()] # 获取当前目录下的所有文件 image_features = {} deleted_count = 0 # 记录删除的图片数量 for file_path in files: if file_path.suffix in {".jpg", ".png"}: # 筛选出图片文件 try: image_feature = extract_image_features(str(file_path)) # 使用str()将Path对象转换为字符串 if image_feature is not None: is_duplicate = False for existing_path, existing_feature in image_features.items(): distance = np.linalg.norm(existing_feature - image_feature) # 计算欧氏距离 if distance < 0.3: # 设定阈值来判断相似度，根据实际情况调整 is_duplicate = True print(f"删除重复图片: {file_path}") file_path.unlink() # 使用unlink()删除文件 deleted_count += 1 break if not is_duplicate: image_features[file_path] = image_feature except Exception as e: print(f"处理图片时出错: {e}") print("已删除 {} 张重复图片".format(deleted_count)) # 加载预训练的ResNet50模型 model = ResNet50(weights='imagenet', include_top=False, pooling='avg') delete_duplicate_images() ``` 最后修改：2024 年 10 月 07 日 © 允许规范转载 赞 如果觉得我的文章对你有用，请随意赞赏