# 19.人脸识别-保存数据库-mysql.py from insightface.app import FaceAnalysis import cv2 import numpy as np import pymysql import os from numpy.linalg import norm import logging # 配置日志 logging.basicConfig( level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s', handlers=[ logging.StreamHandler(), # 输出到控制台 logging.FileHandler('face_recognition.log', encoding='utf-8') # 同时保存到文件 ] ) logger = logging.getLogger(__name__) # 初始化InsightFace模型 app = FaceAnalysis(name='buffalo_l', root='./') app.prepare(ctx_id=0, det_size=(640, 640)) class FaceDatabase: def __init__(self, host='localhost', port=3306, user='root', password='root', database='face_repository'): self.host = host self.port = port self.user = user self.password = password self.database = database self.init_database() # 获取数据库连接 def get_connection(self): return pymysql.connect( host=self.host, port=self.port, user=self.user, password=self.password, database=self.database, charset='utf8mb4' ) # 初始化数据库 def init_database(self): connection = pymysql.connect( host=self.host, port=self.port, user=self.user, password=self.password, charset='utf8mb4' ) # 添加人脸到数据库 def add_face(self, name, embedding): connection = self.get_connection() try: with connection.cursor() as cursor: # 将numpy数组转换为bytes embedding_blob = embedding.tobytes() cursor.execute( 'INSERT INTO faces (name, embedding) VALUES (%s, %s)', (name, embedding_blob) ) connection.commit() logger.info(f"✅ 成功注册: {name}") return True except pymysql.IntegrityError: logger.warning(f"⚠️ 姓名已存在: {name}") return False except Exception as e: logger.error(f"添加人脸失败: {e}") return False finally: connection.close() # 获取所有人脸数据 def get_all_faces(self): connection = self.get_connection() try: with connection.cursor() as cursor: cursor.execute('SELECT name, embedding FROM faces') results = cursor.fetchall() face_database = {} for name, embedding_blob in results: # 将bytes转换回numpy数组 embedding = np.frombuffer(embedding_blob, dtype=np.float32) face_database[name] = embedding logger.info(f"📊 从数据库加载了 {len(face_database)} 个人脸特征") return face_database except Exception as e: logger.error(f"获取人脸数据失败: {e}") return {} finally: connection.close() # 删除人脸 def delete_face(self, face_id): connection = self.get_connection() try: with connection.cursor() as cursor: cursor.execute('DELETE FROM faces WHERE id = %s', (face_id,)) affected_rows = cursor.rowcount connection.commit() if affected_rows > 0: logger.info(f"🗑️ 已删除: {face_id}") return True else: logger.warning(f"❌ 未找到: {face_id}") return False except Exception as e: logger.error(f"删除人脸失败: {e}") return False finally: connection.close() # 检查人脸是否存在 def face_exists(self, name): connection = self.get_connection() try: with connection.cursor() as cursor: cursor.execute('SELECT 1 FROM faces WHERE name = %s', (name,)) exists = cursor.fetchone() is not None return exists except Exception as e: logger.error(f"检查人脸是否存在失败: {e}") return False finally: connection.close() # 创建全局数据库实例 face_db = FaceDatabase(host='localhost', port=3306, user='root', password='root', database='face_repository') # 注册人脸 def register_face(image_path, person_name): # 检查图片是否存在 if not os.path.exists(image_path): logger.error(f"❌ 图片不存在: {image_path}") return False # 读取图片 img = cv2.imread(image_path) if img is None: logger.error(f"❌ 无法读取图片: {image_path}") return False # 使用InsightFace检测人脸 faces = app.get(img) if len(faces) == 0: logger.error(f"❌ 在 {image_path} 中未检测到人脸") return False elif len(faces) > 1: logger.warning(f"⚠️ 在 {image_path} 中检测到 {len(faces)} 张人脸，使用第一张") # 提取人脸特征 embedding = faces[0].embedding logger.info(f"📐 提取到特征向量，维度: {embedding.shape}") # 保存到数据库 return face_db.add_face(person_name, embedding) # 从文件夹批量注册人脸 def register_faces_from_folder(folder_path): if not os.path.exists(folder_path): logger.error(f"❌ 文件夹不存在: {folder_path}") return supported_formats = ('.jpg', '.jpeg', '.png', '.bmp', '.tiff') registered_count = 0 for filename in os.listdir(folder_path): if filename.lower().endswith(supported_formats): # 使用文件名（不含扩展名）作为人名 person_name = os.path.splitext(filename)[0] image_path = os.path.join(folder_path, filename) if register_face(image_path, person_name): registered_count += 1 logger.info(f"🎉 批量注册完成，共注册 {registered_count} 个人脸") # 识别人脸 def recognize_face(image_path, threshold=0.6): # 检查图片是否存在 if not os.path.exists(image_path): logger.error(f"❌ 图片不存在: {image_path}") return # 读取图片 img = cv2.imread(image_path) if img is None: logger.error(f"❌ 无法读取图片: {image_path}") return # 从数据库加载所有人脸特征 face_database = face_db.get_all_faces() if not face_database: logger.error("❌ 数据库中无人脸数据，请先注册人脸") return # 使用InsightFace检测人脸 faces = app.get(img) if len(faces) == 0: logger.error("❌ 未检测到人脸") return logger.info(f"🔍 检测到 {len(faces)} 张人脸，开始识别...") for i, face in enumerate(faces): unknown_embedding = face.embedding max_sim = -1 identity = "未知" best_match_name = None # 与数据库中的每个人脸进行比对 for name, known_embedding in face_database.items(): # 计算余弦相似度 cos_sim = unknown_embedding @ known_embedding.T / (norm(unknown_embedding) * norm(known_embedding)) if cos_sim > max_sim: max_sim = cos_sim identity = name best_match_name = name # 根据阈值决定最终身份 if max_sim < threshold: identity = "未知" # 输出结果 status = "✅" if identity != "未知" else "❓" logger.info(f"{status} 人脸 {i + 1}: 识别为 【{identity}】, 相似度: {max_sim:.4f}") # 显示匹配详情 if identity != "未知": matched_embedding = face_database[best_match_name] logger.info(f" 匹配特征: {best_match_name}, 范数: {norm(matched_embedding):.4f}") # 列出所有已注册的人脸 def list_registered_faces(): face_database = face_db.get_all_faces() if not face_database: logger.info("📭 数据库为空") return logger.info(f"\n📋 已注册的人脸 ({len(face_database)} 个):") for i, name in enumerate(face_database.keys(), 1): logger.info(f" {i}. {name}") # 删除指定人脸 def delete_face(face_id): return face_db.delete_face(face_id) # 使用示例 if __name__ == "__main__": logger.info("=" * 50) logger.info("🎭 InsightFace 人脸识别系统") logger.info("=" * 50) # # 1. 批量注册人脸 # logger.info("\n1.批量注册人脸") # register_faces_from_folder('./resources/face') # # # 2. 单个注册 # logger.info("\n2.单个注册") # register_face('./resources/face/01.png', '张三') # register_face('./resources/face/02.png', '李四') # # # 3. 查看已注册的人脸 # logger.info("\n3.已注册人脸列表") # list_registered_faces() # 4. 识别人脸 logger.info("\n4.人脸识别测试") recognize_face('./resources/face/16.jpg', threshold=0.6) # 5. 删除测试 # logger.info("\n5. 🗑️ 删除人脸测试") # delete_face('测试删除') logger.info("\n" + "=" * 50) logger.info("程序执行完成")