Intelligent-Elderly-Care/Class/detection/fall_detection.py

# # From Python
# # It requires OpenCV installed for Python
import math
import sys
import time
import cv2
import os
from sys import platform
import argparse
import numpy as np

try:
    # Import Openpose (Windows/Ubuntu/OSX)
    # dir_path = os.path.dirname(os.path.realpath(__file__))
    dir_path = 'D:\\BJTU\\Python\\openpose-master\\build-2017'
    try:
        # Windows Import
        if platform == "win32":
            # Change these variables to point to the correct folder (Release/x64 etc.)
            sys.path.append(dir_path + '\\python\\openpose\\Release')
            os.environ['PATH'] = os.environ['PATH'] + ';' + dir_path + '\\x64\\Release;' + dir_path + '\\bin;'
            import pyopenpose as op
        else:
            # Change these variables to point to the correct folder (Release/x64 etc.)
            sys.path.append('../../python')
            # If you run `make install` (default path is `/usr/local/python` for Ubuntu), you can also access the OpenPose/python module from there. This will install OpenPose and the python library at your desired installation path. Ensure that this is in your python path in order to use it.
            # sys.path.append('/usr/local/python')
            from openpose import pyopenpose as op
    except ImportError as e:
        print(
            'Error: OpenPose library could not be found. Did you enable `BUILD_PYTHON` in CMake and have this Python script in the right folder?')
        raise e
    

    # cap = cv2.VideoCapture(0)
    cap = cv2.VideoCapture("demo.mp4")

    _, frame = cap.read()
    cv2.imwrite('fall_detection.jpg', frame)

    # Flags
    parser = argparse.ArgumentParser()
    parser.add_argument("--image_path",
                        default="fall_detection.jpg",
                        help="Process an image. Read all standard formats (jpg, png, bmp, etc.).")
    args = parser.parse_known_args()

    # Custom Params (refer to include/openpose/flags.hpp for more parameters)
    params = dict()
    params["model_folder"] = "D:\\BJTU\\Python\\openpose-master\\models\\"
    params['net_resolution'] = "-1x160"

    # Add others in path?
    for i in range(0, len(args[1])):
        curr_item = args[1][i]
        if i != len(args[1]) - 1:
            next_item = args[1][i + 1]
        else:
            next_item = "1"
        if "--" in curr_item and "--" in next_item:
            key = curr_item.replace('-', '')
            if key not in params:  params[key] = "1"
        elif "--" in curr_item and "--" not in next_item:
            key = curr_item.replace('-', '')
            if key not in params: params[key] = next_item

    # Starting OpenPose
    opWrapper = op.WrapperPython()
    opWrapper.configure(params)
    opWrapper.start()

    # init variables
    frame_start_time = 0
    v0 = 0
    width0 = []
    height0 = []
    center0 = []
    couter = 0
    error = 0

    # tolerance
    lastFrame1 = None
    lastFrame2 = None

    # Construct it from system arguments
    # op.init_argv(args[1])
    # oppython = op.OpenposePython()
    while cap.isOpened():
        flag, frame = cap.read()
        kk = cv2.waitKey(1)
        # 按下 q 键退出
        if kk == ord('q'):
            break
        else:
            cv2.imwrite('fall_detection.jpg', frame)
            height = []
            width = []
            center = []

            # Process Image
            datum = op.Datum()
            imageToProcess = cv2.imread(args[0].image_path)
            datum.cvInputData = imageToProcess
            opWrapper.emplaceAndPop([datum])

            # Display Image
            # print("Body keypoints: \n" + str(datum.poseKeypoints))

            # fall judge
            try:
                # key points have been identified
                x = datum.poseKeypoints[0][:, 0]
                y = datum.poseKeypoints[0][:, 1]

                width.append(np.max(x[np.nonzero(x)]) - np.min(x[np.nonzero(x)]))
                height.append(np.max(y[np.nonzero(y)]) - np.min(y[np.nonzero(y)]))

                center.append(np.mean(x[np.nonzero(x)]))
                center.append(np.mean(y[np.nonzero(y)]))

                if frame_start_time == 0:
                    center0 = center.copy()
                    width0 = width.copy()
                    height0 = height.copy()
                    frame_start_time = time.time()
                else:
                    diff = np.array([center[0] - center0[0], center[1] - center0[1]])
                    dist = math.sqrt(np.sum((diff * 10 ** (-4)) ** 2))
                    now = time.time()
                    v = dist / (now - frame_start_time)
                    a = (v ** 2 - v0 ** 2) / (2 * dist)

                    # print(v, abs(a))
                    if (abs(a) > 0.2) and \
                            (np.subtract(np.array(width), np.array(height)) > np.subtract(np.array(width0),
                                                                                          np.array(height0))):
                        couter += 1
                        # print("alarm by v and a")
                    elif (width > height and (x[8] != 0 or x[9] != 0 or x[12] != 0)):
                        couter += 1
                        # print("alarm by w and h")
                    else:
                        if error == 0:
                            error += 1
                        else:
                            couter = 0
                            error = 0

                    if couter > 2:
                        print("alarm")

                    # update variables
                    frame_start_time = now
                    v0 = v
                    width0 = width.copy()
                    height0 = height.copy()
                # if width > height:
                #     print("alarm")
            except:
                # 如果第一二帧是None，对其进行初始化,计算第一二帧的不同
                if lastFrame2 is None:
                    if lastFrame1 is None:
                        lastFrame1 = frame
                    else:
                        lastFrame2 = frame
                        global frameDelta1  # 全局变量
                        frameDelta1 = cv2.absdiff(lastFrame1, lastFrame2)  # 帧差一
                    continue

                # 计算当前帧和前帧的不同,计算三帧差分
                frameDelta2 = cv2.absdiff(lastFrame2, frame)  # 帧差二
                thresh = cv2.bitwise_and(frameDelta1, frameDelta2)  # 图像与运算
                thresh2 = thresh.copy()

                # 当前帧设为下一帧的前帧,前帧设为下一帧的前前帧,帧差二设为帧差一
                lastFrame1 = lastFrame2
                lastFrame2 = frame.copy()
                frameDelta1 = frameDelta2

                # 结果转为灰度图
                thresh = cv2.cvtColor(thresh, cv2.COLOR_BGR2GRAY)

                # 图像二值化
                thresh = cv2.threshold(thresh, 25, 255, cv2.THRESH_BINARY)[1]

                # 去除图像噪声,先腐蚀再膨胀(形态学开运算)
                thresh = cv2.dilate(thresh, None, iterations=3)
                thresh = cv2.erode(thresh, None, iterations=1)

                # 阀值图像上的轮廓位置
                major = cv2.__version__.split('.')[0]
                if major == '3':
                    binary, cnts, hierarchy = cv2.findContours(thresh.copy(), cv2.RETR_EXTERNAL,
                                                               cv2.CHAIN_APPROX_SIMPLE)
                else:
                    cnts, hierarchy = cv2.findContours(thresh.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

                # 遍历轮廓
                for c in cnts:
                    # 忽略小轮廓，排除误差
                    if cv2.contourArea(c) < 300:
                        continue

                    # 计算轮廓的边界框，在当前帧中画出该框
                    (x, y, w, h) = cv2.boundingRect(c)
                    cv2.rectangle(frame, (x, y), (x + w, y + h), (0, 255, 0), 2)

                # 显示当前帧
                cv2.imshow("OpenPose 1.6.0 - Tutorial Python API", frame)
                continue

            cv2.imshow("OpenPose 1.6.0 - Tutorial Python API", datum.cvOutputData)
    cap.release()
    cv2.destroyAllWindows()
except Exception as e:
    print(e)
    sys.exit(-1)