openpose计算人体骨骼角度

基本原理与实例代码
openpose输出的关键点信息如下图，总共有25个关节点

对应的关节名称如下：

使用的是余弦定理计算角度：

夹角用大写A、B、C表示，边长用小写a、b、c来表示，余弦定理的数学公式如下：

python代码实例：

# keypoints_from_images.pyimport sys
import cv2
import os
from sys import platform
import argparse
import time
import mathdef angle_between_points( p0, p1, p2 ):# 计算角度a = (p1[0]-p0[0])**2 + (p1[1]-p0[1])**2b = (p1[0]-p2[0])**2 + (p1[1]-p2[1])**2c = (p2[0]-p0[0])**2 + (p2[1]-p0[1])**2if a * b == 0:return -1.0 return  math.acos( (a+b-c) / math.sqrt(4*a*b) ) * 180 /math.pidef length_between_points(p0, p1):# 2点之间的距离return math.hypot(p1[0]- p0[0], p1[1]-p0[1])def get_angle_point(human, pos):# 返回各个部位的关键点pnts = []if pos == 'left_elbow':pos_list = (5,6,7)elif pos == 'left_hand':pos_list = (1,5,7)elif pos == 'left_knee':pos_list = (12,13,14)elif pos == 'left_ankle':pos_list = (5,12,14)elif pos == 'right_elbow':pos_list = (2,3,4)elif pos == 'right_hand':pos_list = (1,2,4)elif pos == 'right_knee':pos_list = (9,10,11)elif pos == 'right_ankle':pos_list = (2,9,11)else:print('Unknown  [%s]', pos)return pntsfor i in range(3):if human[pos_list[i]][2] <= 0.1:print('component [%d] incomplete'%(pos_list[i]))return pntspnts.append((int( human[pos_list[i]][0]), int( human[pos_list[i]][1])))return pntsdef angle_left_hand(human):pnts = get_angle_point(human, 'left_hand')if len(pnts) != 3:print('component incomplete')return -1angle = 0if pnts is not None:angle = angle_between_points(pnts[0], pnts[1], pnts[2])print('left hand angle:%f'%(angle))return angledef angle_left_elbow(human):pnts = get_angle_point(human, 'left_elbow')if len(pnts) != 3:print('component incomplete')returnangle = 0if pnts is not None:angle = angle_between_points(pnts[0], pnts[1], pnts[2])print('left elbow angle:%f'%(angle))return angledef angle_left_knee(human):pnts = get_angle_point(human, 'left_knee')if len(pnts) != 3:print('component incomplete')returnangle = 0if pnts is not None:angle = angle_between_points(pnts[0], pnts[1], pnts[2])print('left knee angle:%f'%(angle))return angledef angle_left_ankle(human):pnts = get_angle_point(human, 'left_ankle')if len(pnts) != 3:print('component incomplete')returnangle = 0if pnts is not None:angle = angle_between_points(pnts[0], pnts[1], pnts[2])print('left ankle angle:%f'%(angle))return angledef angle_right_hand(human):pnts = get_angle_point(human, 'right_hand')if len(pnts) != 3:print('component incomplete')returnangle = 0if pnts is not None:angle = angle_between_points(pnts[0], pnts[1], pnts[2])print('right hand angle:%f'%(angle))return angledef angle_right_elbow(human):pnts = get_angle_point(human, 'right_elbow')if len(pnts) != 3:print('component incomplete')returnangle = 0if pnts is not None:angle = angle_between_points(pnts[0], pnts[1], pnts[2])print('right elbow angle:%f'%(angle))return angledef angle_right_knee(human):pnts = get_angle_point(human, 'right_knee')if len(pnts) != 3:print('component incomplete')returnangle = 0if pnts is not None:angle = angle_between_points(pnts[0], pnts[1], pnts[2])print('right knee angle:%f'%(angle))return angledef angle_right_ankle(human):pnts = get_angle_point(human, 'right_ankle')if len(pnts) != 3:print('component incomplete')returnangle = 0if pnts is not None:angle = angle_between_points(pnts[0], pnts[1], pnts[2])print('right ankle angle:%f'%(angle))return angletry:# Import Openpose (Windows/Ubuntu/OSX)dir_path = os.path.dirname(os.path.realpath(__file__))try:# Windows Importif platform == "win32":# Change these variables to point to the correct folder (Release/x64 etc.)sys.path.append(dir_path + '/../../build/python/openpose/Debug');# sys.path.append(dir_path + '/../../build/python/openpose/Debug')# os.environ['PATH']  = os.environ['PATH'] + ';' + dir_path + '/../../build/x64/Debug;' +  dir_path + '/../../bin;'os.environ['PATH']  = os.environ['PATH'] + ';' + dir_path + '/../../build/x64/Debug;' +  dir_path + '/../../build/bin;'import pyopenpose as opelse:# 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 opexcept 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# Flagsparser = argparse.ArgumentParser()parser.add_argument("--image_dir", default="images", help="Process a directory of images. Read all standard formats (jpg, png, bmp, etc.).")parser.add_argument("--no_display", default=False, help="Enable to disable the visual display.")args = parser.parse_known_args()# Custom Params (refer to include/openpose/flags.hpp for more parameters)params = dict()params["model_folder"] = "../../models/"# 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 OpenPoseopWrapper = op.WrapperPython()opWrapper.configure(params)opWrapper.start()# Read frames on directoryimagePaths = op.get_images_on_directory(args[0].image_dir);# Process and display imagesfor imagePath in imagePaths:datum = op.Datum()imageToProcess = cv2.imread(imagePath)datum.cvInputData = imageToProcessopWrapper.emplaceAndPop(op.VectorDatum([datum]))print("Body keypoints: \n" + str(datum.poseKeypoints))human_count = len(datum.poseKeypoints)for i in range(human_count):for j in range(25):print(datum.poseKeypoints[i][j][0])for i in range(human_count):angle_left_hand(datum.poseKeypoints[i] )angle_left_elbow(datum.poseKeypoints[i] )angle_left_knee(datum.poseKeypoints[i] )angle_left_ankle(datum.poseKeypoints[i] )angle_right_hand(datum.poseKeypoints[i] )angle_right_elbow(datum.poseKeypoints[i] )angle_right_knee(datum.poseKeypoints[i] )angle_right_ankle(datum.poseKeypoints[i] )if not args[0].no_display:cv2.imshow("OpenPose 1.7.0 - Tutorial Python API", datum.cvOutputData)cv2.imwrite("{}.jpg".format(time.time()), datum.cvOutputData)key = cv2.waitKey(15)if key == 27: breakexcept Exception as e:print(e)sys.exit(-1)

测试图片，执行代码：

python keypoints_from_images.py --image_dir ..\media

结果

Starting OpenPose Python Wrapper...
Auto-detecting all available GPUs... Detected 1 GPU(s), using 1 of them starting at GPU 0.
left hand angle:72.597296
left elbow angle:105.013366
left knee angle:150.508638
left ankle angle:157.977345
right hand angle:91.741660
right elbow angle:111.680674
right knee angle:154.645155
right ankle angle:174.466738
component [7] incomplete
component incomplete
component [6] incomplete
component incomplete
left knee angle:99.754916
left ankle angle:149.442500
right hand angle:135.324925
right elbow angle:169.808498
component [10] incomplete
component incomplete
component [11] incomplete
component incomplete

以左手为例，它对应的三个关键点是(1, 2, 4)，也就是下图中的三根黑线的交点，计算出来的角度如图所示，为72.597296度：

openpose计算人体骨骼角度相关推荐

Openpose人体骨骼、手势--静态图像标记及分类2（附源码）
文章目录一.前言二.手势 1.分析 2.效果演示 3.完整代码一.前言首先要知道,通过openpose对手势进行识别,需要找到骨骼点 ,才能确定手部和脸部的位置 ,再识别手部和脸部. 前一篇文 ...
姿态估计：人体骨骼关键点检测综述（2016-2020）
点上方蓝字计算机视觉联盟获取更多干货在右上方 ··· 设为星标 ★,与你不见不散仅作学术分享,不代表本公众号立场,侵权联系删除转载于:作者丨七酱@知乎来源丨https://zhuanlan.z ...
人体骨骼检测相关综述及总结
一.简介实现人体检测,通常采用人体姿态估计(Human Posture Estimation),即将图片中已检测到的人体关键点正确联系起来,从而实现人体姿态的估计,实现人体的检测.人体关键点通常对应 ...
人体骨骼关键点检测综述(2)
五.OpenPose库OpenPose人体姿态识别项目是美国卡耐基梅隆大学(CMU)基于卷积神经网络和监督学习并以Caffe为框架开发的开源库.可以实现人体动作.面部表情.手指运动等姿态估计.适用于单 ...
计算机视觉方向简介 | 人体骨骼关键点检测综述
什么是人体骨骼关键点检测人体骨骼关键点检测,即Pose Estimation,主要检测人体的一些关键点,如关节,五官等,通过关键点描述人体骨骼信息: 应用与挑战人体骨骼关键点检测是计算机视觉的基础 ...
java人体识别_用华为HMS ML kit人体骨骼识别技术，Android快速实现人体姿势动作抓拍...
items = results.getAnalyseList(); // 开发者根据需要处理识别结果,例如,在此方法中进行相似度计算,从而在检测到特定姿势后进行拍照等操作. // 需要注意,这里只对检 ...
人体骨骼关键点检测综述
原创声明:本文为 SIGAI 原创文章,仅供个人学习使用,未经允许,不得转载,不能用于商业目的. 本文及其它机器学习.深度学习算法的全面系统讲解可以阅读<机器学习与应用>,清华大学出版社, ...
计算机视觉人体骨骼点动作识别-1.训练自己的关键点检测模型
人体关键点检测算法关键点并不特指人体骨骼关键点,还有人脸关键点,物体的关键点.其中人体的关键点,也叫作pose Estimation,是最热门,也是最有难度,应用最广的. 应用可以包括:行为识别,人 ...
不需代码 open pose多人动态人体骨骼识别，面部，手部，姿势识别
最近在学open pose相关的东西,OpenPose是卡内基梅隆大学的开源项目,它能支持2D和3D的多人关键点识别,支持手部.面部.脚部识别,单人运动姿势估计. 用官方提供的案例实现识别图片和视频中 ...

openpose计算人体骨骼角度

openpose计算人体骨骼角度相关推荐

最新文章

热门文章