通過Opencv+Openpose實(shí)現(xiàn)體姿態(tài)檢測(cè)

通過一個(gè)偶然機(jī)會(huì)，我了解到了人體姿態(tài)解算，在學(xué)習(xí)K210之余，我便想著通過opencv實(shí)現(xiàn)這個(gè)功能，查找了很多資料，發(fā)現(xiàn)可以利用opencv+openpose實(shí)現(xiàn)，接著我又開始找一些資料，在pycharm上部署。

前言

人體姿態(tài)估計(jì)的一個(gè)有趣應(yīng)用是 CGI（computer graphic image，一種電影制造技術(shù)）應(yīng)用。如果可以檢測(cè)出人體姿態(tài)，那么圖形、風(fēng)格、特效增強(qiáng)、設(shè)備和藝術(shù)造型等就可以被加載在人體上。

通過追蹤人體姿態(tài)的變化，渲染的圖形可以在人動(dòng)的時(shí)候“自然”地與人“融合”。姿態(tài)估計(jì)的一個(gè)有趣應(yīng)用是在交互游戲中追蹤人體對(duì)象的運(yùn)動(dòng)。

比較流行的 Kinect 使用 3D 姿態(tài)估計(jì)（采用 IR 傳感器數(shù)據(jù)）來(lái)追蹤人類玩家的運(yùn)動(dòng)，從而利用它來(lái)渲染虛擬人物的動(dòng)作。

應(yīng)用：

用于檢測(cè)一個(gè)人是否摔倒或疾病

用于健身、體育和舞蹈等的自動(dòng)教學(xué)

用于理解全身的肢體語(yǔ)言（如機(jī)場(chǎng)跑道信號(hào)、交警信號(hào)等）

用于增強(qiáng)安保和監(jiān)控

一、環(huán)境配置

pycharm2021.2.2

pycharm是一個(gè)很好用的軟件，剛開始我們必須要配置相應(yīng)的環(huán)境，當(dāng)然你使用我主頁(yè)里那篇模型訓(xùn)練的環(huán)境也可以，在你運(yùn)行的時(shí)候系統(tǒng)會(huì)提示你缺少了什么環(huán)境，并讓你安裝，你直接安裝即可。這里我就不過多的贅述了。

1.導(dǎo)入文件

在pycharm上導(dǎo)入相應(yīng)的文件后，你可以直接點(diǎn)擊運(yùn)行，系統(tǒng)會(huì)提示你缺少了什么環(huán)境，缺少什么就安裝什么，通過終端使用pip安裝即可。

需要的留下郵箱即可，我發(fā)給你。

2.具體代碼

# To use Inference Engine backend, specify location of plugins:
# export LD_LIBRARY_PATH=/opt/intel/deeplearning_deploymenttoolkit/deployment_tools/external/mklml_lnx/lib:$LD_LIBRARY_PATH
import cv2 as cv
import numpy as np
import argparse


parser = argparse.ArgumentParser()
parser.add_argument('--input', help='Path to image or video. Skip to capture frames from camera')
parser.add_argument('--thr', default=0.2, type=float, help='Threshold value for pose parts heat map')
parser.add_argument('--width', default=368, type=int, help='Resize input to specific width.')
parser.add_argument('--height', default=368, type=int, help='Resize input to specific height.')


args = parser.parse_args()


BODY_PARTS = { "Nose": 0, "Neck": 1, "RShoulder": 2, "RElbow": 3, "RWrist": 4,
        "LShoulder": 5, "LElbow": 6, "LWrist": 7, "RHip": 8, "RKnee": 9,
        "RAnkle": 10, "LHip": 11, "LKnee": 12, "LAnkle": 13, "REye": 14,
        "LEye": 15, "REar": 16, "LEar": 17, "Background": 18 }


POSE_PAIRS = [ ["Neck", "RShoulder"], ["Neck", "LShoulder"], ["RShoulder", "RElbow"],
        ["RElbow", "RWrist"], ["LShoulder", "LElbow"], ["LElbow", "LWrist"],
        ["Neck", "RHip"], ["RHip", "RKnee"], ["RKnee", "RAnkle"], ["Neck", "LHip"],
        ["LHip", "LKnee"], ["LKnee", "LAnkle"], ["Neck", "Nose"], ["Nose", "REye"],
        ["REye", "REar"], ["Nose", "LEye"], ["LEye", "LEar"] ]


inWidth = args.width
inHeight = args.height


net = cv.dnn.readNetFromTensorflow("graph_opt.pb")


cap = cv.VideoCapture(args.input if args.input else 0)


while cv.waitKey(1) < 0:
 ? ?hasFrame, frame = cap.read()
 ? ?if not hasFrame:
 ? ? ? ?cv.waitKey()
 ? ? ? ?break


 ? ?frameWidth = frame.shape[1]
 ? ?frameHeight = frame.shape[0]
 ? ?
 ? ?net.setInput(cv.dnn.blobFromImage(frame, 1.0, (inWidth, inHeight), (127.5, 127.5, 127.5), swapRB=True, crop=False))
 ? ?out = net.forward()
 ? ?out = out[:, :19, :, :] ?# MobileNet output [1, 57, -1, -1], we only need the first 19 elements


 ? ?assert(len(BODY_PARTS) == out.shape[1])


 ? ?points = []
 ? ?for i in range(len(BODY_PARTS)):
 ? ? ? ?# Slice heatmap of corresponging body's part.
 ? ? ? ?heatMap = out[0, i, :, :]


 ? ? ? ?# Originally, we try to find all the local maximums. To simplify a sample
 ? ? ? ?# we just find a global one. However only a single pose at the same time
 ? ? ? ?# could be detected this way.
 ? ? ? ?_, conf, _, point = cv.minMaxLoc(heatMap)
 ? ? ? ?x = (frameWidth * point[0]) / out.shape[3]
 ? ? ? ?y = (frameHeight * point[1]) / out.shape[2]
 ? ? ? ?# Add a point if it's confidence is higher than threshold.
 ? ? ? ?points.append((int(x), int(y)) if conf > args.thr else None)


  for pair in POSE_PAIRS:
    partFrom = pair[0]
    partTo = pair[1]
    assert(partFrom in BODY_PARTS)
    assert(partTo in BODY_PARTS)


    idFrom = BODY_PARTS[partFrom]
    idTo = BODY_PARTS[partTo]


    if points[idFrom] and points[idTo]:
      cv.line(frame, points[idFrom], points[idTo], (0, 255, 0), 3)
      cv.ellipse(frame, points[idFrom], (3, 3), 0, 0, 360, (0, 0, 255), cv.FILLED)
      cv.ellipse(frame, points[idTo], (3, 3), 0, 0, 360, (0, 0, 255), cv.FILLED)


  t, _ = net.getPerfProfile()
  freq = cv.getTickFrequency() / 1000
  cv.putText(frame, '%.2fms' % (t / freq), (10, 20), cv.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 0))


  cv.imshow('OpenPose using OpenCV', frame)

這里便是主函數(shù)的代碼。

3.效果展示

這副圖片便是識(shí)別的效果，幀率還是很不錯(cuò)的。

三、效果優(yōu)化

這個(gè)幀率雖然可以，但是效果屬實(shí)有點(diǎn)拉跨。教我K210的學(xué)長(zhǎng)便指導(dǎo)我進(jìn)行優(yōu)化改進(jìn)，這里附上學(xué)長(zhǎng)的連接

1.具體代碼

import cv2
import time
import mediapipe as mp
from tqdm import tqdm


# 導(dǎo)入solution
mp_pose = mp.solutions.pose




mp_drawing = mp.solutions.drawing_utils




pose = mp_pose.Pose(static_image_mode=False,
      #  model_complexity=1,
          smooth_landmarks=True,
      #    enable_segmentation=True,
          min_detection_confidence=0.5,
          min_tracking_confidence=0.5)




def process_frame(img):
  # BGR轉(zhuǎn)RGB
  img_RGB = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)


  results = pose.process(img_RGB)


  # 可視化
  mp_drawing.draw_landmarks(img, results.pose_landmarks, mp_pose.POSE_CONNECTIONS)
  # look_img(img)


  # mp_drawing.plot_landmarks(results.pose_world_landmarks, mp_pose.POSE_CONNECTIONS)


  #   # BGR轉(zhuǎn)RGB
  #   img_RGB = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
  #
  #   results = hands.process(img_RGB)


  #   if results.multi_hand_landmarks: # 如果有檢測(cè)到手
  #
  #     for hand_idx in range(len(results.multi_hand_landmarks)):
  #       hand_21 = results.multi_hand_landmarks[hand_idx]
  #       mpDraw.draw_landmarks(img, hand_21, mp_hands.HAND_CONNECTIONS)


  return img


cap = cv2.VideoCapture(1)


# 打開cap
cap.open(0)


# 無(wú)限循環(huán)，直到break被觸發(fā)
while cap.isOpened():
  # 獲取畫面
  success, frame = cap.read()
  if not success:
    print('Error')
    break


  ## !!!處理幀函數(shù)
  frame = process_frame(frame)


  # 展示處理后的三通道圖像
  cv2.imshow('my_window', frame)


  if cv2.waitKey(1) in [ord('q'), 27]:
    break




cap.release()




cv2.destroyAllWindows()

2.效果展示

效果簡(jiǎn)直太好了

總結(jié)

到這里這篇文章就結(jié)束了，寫這篇博客只是單純記錄自己的學(xué)習(xí)過程。希望看到這篇博客的你，能夠更加堅(jiān)定的學(xué)習(xí)。胡適說(shuō)過一句話我覺得特別好，這里分享給大家。

怕什么真理無(wú)窮，進(jìn)一寸有進(jìn)一寸的歡喜。

審核編輯：湯梓紅