OpenCV+Python+Dlib面部标定、眨眼和疲劳检测代码

以下是OpenCV、Python和Dlib实现面部标定、眨眼和疲劳检测的代码：

首先，需要安装OpenCV、Python和Dlib库。在Python中，可以使用pip安装这些库：

pip install opencv-python
pip install dlib

接下来，导入必要的库：

import cv2
import dlib
import numpy as np
from scipy.spatial import distance as dist

然后，定义一些常量和函数：

EYE_AR_THRESH = 0.25  # 眼睛长宽比阈值
EYE_AR_CONSEC_FRAMES = 3  # 连续帧数
YAWN_THRESH = 20  # 打哈欠阈值
ALARM_SOUND_PATH = "alarm.wav"  # 警报声音文件路径

def eye_aspect_ratio(eye):
    # 计算眼睛长宽比
    A = dist.euclidean(eye[1], eye[5])
    B = dist.euclidean(eye[2], eye[4])
    C = dist.euclidean(eye[0], eye[3])
    ear = (A + B) / (2.0 * C)
    return ear

def mouth_aspect_ratio(mouth):
    # 计算嘴巴长宽比
    A = dist.euclidean(mouth[14], mouth[18])
    B = dist.euclidean(mouth[12], mouth[16])
    C = dist.euclidean(mouth[0], mouth[6])
    mar = (A + B) / (2.0 * C)
    return mar

def play_alarm_sound(path):
    # 播放警报声音
    import os
    os.system("aplay " + path + " &")

现在，让我们加载Dlib的人脸检测器和68个面部标定点模型：

detector = dlib.get_frontal_face_detector()
predictor = dlib.shape_predictor("shape_predictor_68_face_landmarks.dat")

最后，我们可以开始处理视频流或摄像头输入：

cap = cv2.VideoCapture(0)  # 摄像头输入
ear_history = []  # 眼睛长宽比历史记录
mar_history = []  # 嘴巴长宽比历史记录
alarm_on = False  # 是否播放警报声音
while True:
    ret, frame = cap.read()
    if not ret:
        break

    gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
    rects = detector(gray, 0)

    for rect in rects:
        # 检测人脸并标定面部
        shape = predictor(gray, rect)
        shape = np.array([(p.x, p.y) for p in shape.parts()])

        # 计算眼睛长宽比和嘴巴长宽比
        left_eye = shape[36:42]
        right_eye = shape[42:48]
        mouth = shape[48:68]
        ear = (eye_aspect_ratio(left_eye) + eye_aspect_ratio(right_eye)) / 2.0
        mar = mouth_aspect_ratio(mouth)

        # 在视频中显示眼睛和嘴巴区域
        cv2.drawContours(frame, [cv2.convexHull(left_eye)], -1, (0, 255, 0), 1)
        cv2.drawContours(frame, [cv2.convexHull(right_eye)], -1, (0, 255, 0), 1)
        cv2.drawContours(frame, [cv2.convexHull(mouth)], -1, (0, 255, 0), 1)

        # 更新眼睛长宽比历史记录
        ear_history.append(ear)
        if len(ear_history) > EYE_AR_CONSEC_FRAMES:
            ear_history.pop(0)

        # 更新嘴巴长宽比历史记录
        mar_history.append(mar)
        if len(mar_history) > EYE_AR_CONSEC_FRAMES:
            mar_history.pop(0)

        # 判断是否眨眼和打哈欠
        ear_avg = np.mean(ear_history)
        mar_avg = np.mean(mar_history)
        if ear_avg < EYE_AR_THRESH and mar_avg > YAWN_THRESH:
            # 开始播放警报声音
            if not alarm_on:
                alarm_on = True
                play_alarm_sound(ALARM_SOUND_PATH)
        else:
            # 停止播放警报声音
            alarm_on = False

    cv2.imshow("Frame", frame)
    key = cv2.waitKey(1)
    if key == 27:
        break

cap.release()
cv2.destroyAllWindows()

这段代码中，我们使用OpenCV从摄像头或视频流中读取帧。然后，我们将每个帧转换为灰度图像，并使用Dlib检测人脸和面部标定点。

接下来，我们计算眼睛长宽比和嘴巴长宽比，并绘制出眼睛和嘴巴的轮廓。然后，我们更新眼睛长宽比和嘴巴长宽比历史记录，并检查是否有连续的帧满足眨眼和打哈欠的条件。如果是，我们播放警报声音。

最后，我们在视频中显示帧，并等待按下ESC键退出程序。

希望这个例子可以帮助你开始使用OpenCV、Python和Dlib实现面部标定、眨眼和疲劳检测。