match_face/scanf_face.py

import cv2
import face_recognition
import os
import sqlite3
import numpy as np
from datetime import datetime
import time

# 初始化摄像头
cap = cv2.VideoCapture(0)
max_photos = 10

# 创建目录以保存照片
save_path = "./captured_faces"
os.makedirs(save_path, exist_ok=True)

def create_face_database(db_name="face_database.db"):
    """创建人脸数据库和匹配日志表"""
    conn = sqlite3.connect(db_name)
    c = conn.cursor()
    c.execute('''CREATE TABLE IF NOT EXISTS faces
                 (id INTEGER PRIMARY KEY AUTOINCREMENT,
                 name TEXT NOT NULL,
                 identity TEXT NOT NULL,
                 image_path TEXT NOT NULL,
                 encoding BLOB NOT NULL)''')

    c.execute('''CREATE TABLE IF NOT EXISTS match_logs
                 (id INTEGER PRIMARY KEY AUTOINCREMENT,
                 name TEXT NOT NULL,
                 identity TEXT NOT NULL,
                 image_path TEXT NOT NULL,
                 match_time TEXT NOT NULL)''')
    conn.commit()
    conn.close()

def add_face_to_database(name, identity, image_path, db_name="face_database.db"):
    """将人脸信息添加到数据库"""
    conn = sqlite3.connect(db_name)
    c = conn.cursor()

    # 将图片路径转换为相对路径
    relative_image_path = os.path.relpath(image_path, start='./static')

    image = face_recognition.load_image_file(image_path)
    face_encodings = face_recognition.face_encodings(image)

    if face_encodings:
        face_encoding = face_encodings[0]
        encoding_blob = np.array(face_encoding).tobytes()
        c.execute("INSERT INTO faces (name, identity, image_path, encoding) VALUES (?, ?, ?, ?)",
                  (name, identity, relative_image_path, encoding_blob))
        conn.commit()
    conn.close()



def match_faces(captured_images, db_name="face_database.db", tolerance=0.4, log_file="match_log.txt"):
    """比对抓拍的图片与数据库中的已知人脸"""
    conn = sqlite3.connect(db_name)
    c = conn.cursor()

    c.execute("SELECT name, identity, image_path, encoding FROM faces")
    known_faces = c.fetchall()

    for image_path in captured_images:
        unknown_image = face_recognition.load_image_file(image_path)
        face_encodings = face_recognition.face_encodings(unknown_image)

        if len(face_encodings) == 0:
            print(f"没有检测到人脸：{image_path}")
            continue

        unknown_encoding = face_encodings[0]

        for name, identity, db_image_path, encoding_blob in known_faces:
            known_encoding = np.frombuffer(encoding_blob, dtype=np.float64)
            match = face_recognition.compare_faces([known_encoding], unknown_encoding, tolerance=tolerance)

            if match[0]:
                print(f"匹配成功: {name} ({identity}) 在 {image_path}")
                log_match(name, identity, db_image_path, db_name, log_file)  # 使用数据库中的 image_path
                # 不更新数据库中的 image_path，因为我们只在匹配时使用它
                conn.commit()
                conn.close()
                return True

        print(f"未发现匹配: 在 {image_path} 中的任何已知人脸")

    conn.close()
    return False



def log_match(name, identity, db_image_path, db_name, log_file):
    """记录匹配结果，将数据库中的图片路径和匹配时间添加到匹配记录表单中"""
    conn = sqlite3.connect(db_name)
    c = conn.cursor()

    # 记录到日志文件
    with open(log_file, 'a') as log:
        log.write(f"{name},{identity},{db_image_path}\n")

    # 将匹配信息插入到匹配日志表
    c.execute("INSERT INTO match_logs (name, identity, image_path, match_time) VALUES (?, ?, ?, ?)",
              (name, identity, db_image_path, datetime.now().strftime("%Y-%m-%d %H:%M:%S")))

    conn.commit()
    conn.close()


# 创建人脸数据库
create_face_database()

# 向数据库中添加人脸
add_face_to_database("李四", "居民", "./static/db_image/test1.jpg")
add_face_to_database("张三", "居民", "./static/db_image/test2.jpg")
add_face_to_database("王五", "居民", "./static/db_image/test3.jpg")

# 主程序循环
while True:
    ret, frame = cap.read()
    if not ret:
        break

    rgb_frame = frame[:, :, ::-1]
    face_locations = face_recognition.face_locations(rgb_frame)

    if face_locations:
        print("检测到人脸，开始抓拍...")

        captured_images = []
        photo_count = 0

        while photo_count < max_photos:
            ret, frame = cap.read()
            if not ret:
                break

            rgb_frame = frame[:, :, ::-1]
            face_locations = face_recognition.face_locations(rgb_frame)

            for face_location in face_locations:
                top, right, bottom, left = face_location
                cv2.rectangle(frame, (left, top), (right, bottom), (0, 255, 0), 2)

                face_image = frame[top:bottom, left:right]
                image_path = os.path.join(save_path, f"face_{photo_count + 1}.jpg")
                cv2.imwrite(image_path, face_image)
                captured_images.append(image_path)

                photo_count += 1
                if photo_count >= max_photos:
                    break

            cv2.imshow("Capturing Faces", frame)
            if cv2.waitKey(1) & 0xFF == ord('q'):
                break

        cv2.destroyAllWindows()

        if match_faces(captured_images):
            print("匹配成功")
        else:
            print("没有匹配")

        print("等待30秒后继续...")
        time.sleep(30)

cap.release()