本篇文章适合刚入门OpenCV的同学们。文章将介绍如何使用Python利用OpenCV图像捕捉，配合强大的Mediapipe库来实现手势检测与识别；本系列后续还会继续更新Mediapipe手势的各种衍生项目，还请多多关注！
视频捕捉帧数稳定在（25-30）
项目的实现，核心是强大的Mediapipe ，它是google的一个开源项目：
| 功能 | 详细 |
|---|---|
| 人脸检测 FaceMesh | 从图像/视频中重建出人脸的3D Mesh |
| 人像分离 | 从图像/视频中把人分离出来 |
| 手势跟踪 | 21个关键点的3D坐标 |
| 人体3D识别 | 33个关键点的3D坐标 |
| 物体颜色识别 | 可以把头发检测出来，并图上颜色 |
Mediapipe Dev
以上是Mediapipe的几个常用功能 ，这几个功能我们会在后续一一讲解实现
Python安装Mediapipe
pip install mediapipe==0.8.9.1
也可以用 setup.py 安装
https://github.com/google/mediapipe
Python 3.7
Mediapipe 0.8.9.1
Numpy 1.21.6
OpenCV-Python 4.5.5.64
OpenCV-contrib-Python 4.5.5.64
实测也支持Python3.8-3.9
OpenCV摄像头捕捉部分：
import cv2
cap = cv2.VideoCapture(0) #OpenCV摄像头调用：0=内置摄像头（笔记本） 1=USB摄像头-1 2=USB摄像头-2
while True:
success, img = cap.read()
imgRGB = cv2.cvtColor(img, cv2.COLOR_BGR2RGB) #cv2图像初始化
cv2.imshow("HandsImage", img) #CV2窗体
cv2.waitKey(1) #关闭窗体
mediapipe 手势识别与绘制
#定义并引用mediapipe中的hands模块
mpHands = mp.solutions.hands
hands = mpHands.Hands()
mpDraw = mp.solutions.drawing_utils
while True:
success, img = cap.read()
imgRGB = cv2.cvtColor(img, cv2.COLOR_BGR2RGB) #cv2图像初始化
results = hands.process(imgRGB)
# print(results.multi_hand_landmarks)
if results.multi_hand_landmarks:
for handLms in results.multi_hand_landmarks:
for id, lm in enumerate(handLms.landmark):
# print(id, lm)
h, w, c = img.shape
cx, cy = int(lm.x * w), int(lm.y * h)
print(id, cx, cy)
# if id == 4:
cv2.circle(img, (cx, cy), 15, (255, 0, 255), cv2.FILLED)
#绘制手部特征点：
mpDraw.draw_landmarks(img, handLms, mpHands.HAND_CONNECTIONS)
import time
#帧率时间计算
pTime = 0
cTime = 0
while True
cTime = time.time()
fps = 1 / (cTime - pTime)
pTime = cTime
cv2.putText(img, str(int(fps)), (10, 70), cv2.FONT_HERSHEY_PLAIN, 3,
(255, 0, 255), 3) #FPS的字号，颜色等设置
# Coding BIGBOSSyifi
# Datatime:2022/4/24 21:41
# Filename:HandsDetector.py
# Toolby: PyCharm
import cv2
import mediapipe as mp
import time
cap = cv2.VideoCapture(0) #OpenCV摄像头调用：0=内置摄像头（笔记本） 1=USB摄像头-1 2=USB摄像头-2
#定义并引用mediapipe中的hands模块
mpHands = mp.solutions.hands
hands = mpHands.Hands()
mpDraw = mp.solutions.drawing_utils
#帧率时间计算
pTime = 0
cTime = 0
while True:
success, img = cap.read()
imgRGB = cv2.cvtColor(img, cv2.COLOR_BGR2RGB) #cv2图像初始化
results = hands.process(imgRGB)
# print(results.multi_hand_landmarks)
if results.multi_hand_landmarks:
for handLms in results.multi_hand_landmarks:
for id, lm in enumerate(handLms.landmark):
# print(id, lm)
h, w, c = img.shape
cx, cy = int(lm.x * w), int(lm.y * h)
print(id, cx, cy)
# if id == 4:
cv2.circle(img, (cx, cy), 15, (255, 0, 255), cv2.FILLED)
#绘制手部特征点：
mpDraw.draw_landmarks(img, handLms, mpHands.HAND_CONNECTIONS)
'''''
视频FPS计算
'''
cTime = time.time()
fps = 1 / (cTime - pTime)
pTime = cTime
cv2.putText(img, str(int(fps)), (10, 70), cv2.FONT_HERSHEY_PLAIN, 3,
(255, 0, 255), 3) #FPS的字号，颜色等设置
cv2.imshow("HandsImage", img) #CV2窗体
cv2.waitKey(1) #关闭窗体
以此篇文章技术为基础，后续会更新利用此篇基础技术实现的《手势控制：音量，鼠标》
项目下载地址https://github.com/BIGBOSS-dedsec/HandsDetection_Python