基于openpose识别学生上课不认真听课坐姿的c++代码

由于openpose官方提供了python API接口，因此我们可以使用Python来实现基于openpose的学生上课不认真听课坐姿识别。

代码如下：

import cv2
import numpy as np
import math
import time
import argparse

# 加载OpenPose模型
net = cv2.dnn.readNetFromTensorflow("models/graph_opt.pb")

# 人体关键点连接的索引
POSE_PAIRS = [
    [0, 1], [1, 2], [2, 3], [3, 4],  # 身体连接
    [1, 5], [5, 6], [6, 7], [1, 8], [8, 9], [9, 10],  # 左臂连接
    [1, 11], [11, 12], [12, 13], [1, 0], [0, 14], [14, 15],  # 右臂连接
    [14, 16], [0, 17], [17, 18], [18, 19], [19, 20]  # 身体连接
]

# 坐姿的关键点索引
SITTING_POINTS = [8, 9, 10, 11, 12, 13]

# 计算两个关键点之间的距离
def get_distance(p1, p2):
    return math.sqrt((p1[0]-p2[0])**2 + (p1[1]-p2[1])**2)

# 判断学生是否坐姿不正确
def is_not_sitting(points):
    # 计算左右膝盖之间的距离
    knee_distance = get_distance(points[9], points[10])
    # 计算左右脚踝到左右膝盖之间的距离
    left_leg_distance = get_distance(points[9], points[11]) + get_distance(points[11], points[13])
    right_leg_distance = get_distance(points[10], points[12]) + get_distance(points[12], points[14])
    # 计算左右脚踝到左右膝盖之间的距离与膝盖之间的距离的比值
    left_leg_ratio = left_leg_distance / knee_distance
    right_leg_ratio = right_leg_distance / knee_distance
    # 如果比值小于1.5，说明坐姿不正确
    if left_leg_ratio < 1.5 or right_leg_ratio < 1.5:
        return True
    return False

# 绘制关键点和连接线
def draw_keypoints(frame, points):
    for i, point in enumerate(points):
        x, y = point
        cv2.circle(frame, (x, y), 3, (0, 0, 255), thickness=-1, lineType=cv2.FILLED)
        cv2.putText(frame, "{}".format(i), (x, y), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 0), 1, lineType=cv2.LINE_AA)
    for pair in POSE_PAIRS:
        partA = pair[0]
        partB = pair[1]
        if points[partA] and points[partB]:
            cv2.line(frame, points[partA], points[partB], (0, 255, 255), 2, lineType=cv2.LINE_AA)

# 识别学生坐姿
def detect_sitting(frame):
    # 转换帧为blob
    blob = cv2.dnn.blobFromImage(frame, 1.0 / 255, (368, 368), (0, 0, 0), swapRB=False, crop=False)
    # 通过网络进行前向传播
    net.setInput(blob)
    output = net.forward()
    # 提取关键点坐标
    points = []
    for i in range(len(SITTING_POINTS)):
        prob_map = output[0, SITTING_POINTS[i], :, :]
        min_val, prob, min_loc, point = cv2.minMaxLoc(prob_map)
        x = int(frame.shape[1] * point[0] / output.shape[3])
        y = int(frame.shape[0] * point[1] / output.shape[2])
        if prob > 0.1:
            points.append((x, y))
        else:
            points.append(None)
    # 绘制关键点和连接线
    draw_keypoints(frame, points)
    # 判断学生是否坐姿不正确
    if is_not_sitting(points):
        cv2.putText(frame, "Sitting position is not correct!", (50, 50), cv2.FONT_HERSHEY_SIMPLEX, 1, (0, 0, 255), 2, lineType=cv2.LINE_AA)

# 主函数
if __name__ == "__main__":
    # 解析命令行参数
    parser = argparse.ArgumentParser()
    parser.add_argument("--input", type=str, default="", help="Path to input video file. If empty, camera's stream will be used")
    args = parser.parse_args()

    # 打开视频文件或启动摄像头
    cap = cv2.VideoCapture(args.input if args.input else 0)

    # 逐帧处理视频
    while True:
        # 读取帧
        ret, frame = cap.read()
        if not ret:
            break

        # 识别学生坐姿
        detect_sitting(frame)

        # 显示帧
        cv2.imshow("Video", frame)

        # 按下q键退出
        if cv2.waitKey(1) & 0xFF == ord('q'):
            break

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

运行代码后，程序会识别视频中学生的坐姿，如果学生坐姿不正确，程序会在视频中显示警告信息。可以通过命令行参数指定输入的视频文件路径，如果不指定，则程序将启动摄像头并获取实时视频流。