Python-Gesichtserkennung basierend auf opencv und face_recognition

Python-Gesichtserkennung basierend auf opencv und face_recognition

Code wie folgt anzeigen:

Verwenden Sie Fotos, um die Fotos zu speichern, die Sie identifizieren müssen. Nur ein Foto pro Person reicht aus.

Registrieren Sie dann den Benutzer über den folgenden Code und starten Sie dann das Programm. Die Kamera wird zur Erkennung aufgerufen.
AddPhoto("Fa Ge", "./photos/fage.jpg")
AddPhoto("华泽", "./photos/huazai.jpg")

# 7人脸检测 和 8人脸识别：7只能检测到人脸，8还可以识别出是谁
import face_recognition
import cv2
import numpy as np
from PIL import Image, ImageDraw, ImageFont

known_face_names = []
known_face_encodings = []


def IsDuplicateName(name):
    if name in known_face_names:
        return True
    return False


def AddPhoto(name, filename):
    image = face_recognition.load_image_file(filename)
    # 用 128 维的向量表示 1 张人脸
    face_encoding = face_recognition.face_encodings(image)
    if len(face_encoding) != 1:
        return False
    known_face_encodings.insert(0, face_encoding[0])
    known_face_names.insert(0, name)
    return True


def PutCNText(image, strs, local, sizes, colour):
    """
    在 frame 中添加文字
    """
    cv2img = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
    pilimg = Image.fromarray(cv2img)
    draw = ImageDraw.Draw(pilimg)
    font = ImageFont.truetype("./simhei.ttf", sizes, encoding="utf-8")
    draw.text(local, strs, colour, font=font)
    return cv2.cvtColor(np.array(pilimg), cv2.COLOR_RGB2BGR)


def FaceRecognition(frame):
    # 尺寸缩放为原来的 1/4，参数 (0, 0) 原意表示输出图像的大小
    # 当指定为 (0, 0) 时，输出图像的大小会根据 fx 和 fy 参数进行计算
    # 缩小图像 4 倍是为了加速人脸检测过程
    small_frame = cv2.resize(frame, (0, 0), fx=0.25, fy=0.25)

    # np.ascontiguousarray() 确保数组在内存中的存储是连续的
    rgb_small_frame = np.ascontiguousarray(small_frame[:, :, ::-1])

    # 得到检测到的人脸位置信息 face_locations
    face_locations = face_recognition.face_locations(rgb_small_frame)

    # 对 rgb_small_frame 中人脸进行编码，得到人脸向量 face_encodings
    # 这个编码信息将用于后续的人脸比对和识别
    face_encodings = face_recognition.face_encodings(rgb_small_frame, face_locations)

    face_names = []
    for face_encoding in face_encodings:
        matches = face_recognition.compare_faces(known_face_encodings, face_encoding, tolerance=0.4)
        # known_face_encodings 为列表，相同人脸为 True
        name = ""

        if True in matches:
            first_match_index = matches.index(True)
            name = known_face_names[first_match_index]
        else:
            name='未录入人员'

        face_names.append(name)

    for (top, right, bottom, left), name in zip(face_locations, face_names):
        top *= 4
        right *= 4
        bottom *= 4
        left *= 4
        if  name =='未录入人员':
              cv2.rectangle(frame, (left, top), (right, bottom), (0, 0, 255), 2)
              frame = PutCNText(frame, name, (left + 6, bottom - 24), 20, (0, 255, 255))
        else:


            cv2.rectangle(frame, (left, top), (right, bottom), (0, 255, 0), 2)
            frame = PutCNText(frame, name, (left + 6, bottom - 24), 20, (255, 255, 255))

    return frame


if __name__ == "__main__":
    # 注册缓存 人名 和 人脸向量，用于后续人脸识别
    AddPhoto("发哥", "./photos/fage.jpg")
    AddPhoto("华仔", "./photos/huazai.jpg")



    video_capture = cv2.VideoCapture(0)
    # 设置视频帧的 宽度 和 高度
    # video_capture.set(cv2.CAP_PROP_FRAME_WIDTH, 1024) # 3
    # video_capture.set(cv2.CAP_PROP_FRAME_HEIGHT, 768) # 4
    # video_capture.set(cv2.CAP_PROP_FOURCC, cv2.VideoWriter_fourcc("M", "J", "P", "G"))

    while video_capture.isOpened():
        # 读取 1 帧视频图像
        ret, frame = video_capture.read()
        # print("frame.shape:", frame.shape)  # frame.shape = (720, 1280, 3) 就是一张图片

        # 如果读取失败，进入下一循环
        if ret == False:
            continue

        frame = FaceRecognition(frame)
        cv2.imshow("Face Recognition", frame)

        # 退出条件
        if cv2.waitKey(1) & 0xFF == ord("q"):
            break

    # 释放资源
    video_capture.release()
    cv2.destroyAllWindows()