【V1.1】基于树莓派的OpenCV-Python摄像头人脸追踪系统

import cv2 import threading import RPi.GPIO as GPIO # import time GPIO.setwarnings(False) GPIO.setmode(GPIO.BCM) GPIO.setup(23, GPIO.OUT) GPIO.setup(24, GPIO.OUT) GPIO.output(23, GPIO.HIGH) GPIO.output(24, GPIO.HIGH) cap = cv2.VideoCapture(0) # 开启摄像头 classifier = cv2.CascadeClassifier('haarcascade_frontalface_alt2.xml') ok, faceImg = cap.read() # 读取摄像头图像 if ok is False: print('无法读取到摄像头！') high=faceImg.shape[0] width=faceImg.shape[1] left_point = width/2+25 right_point = width/2-25 gray = cv2.cvtColor(faceImg,cv2.COLOR_BGR2GRAY) faceRects = classifier.detectMultiScale(gray,scaleFactor=1.2,minNeighbors=3,minSize=(32, 32)) close=0 def LEFT(): GPIO.output(23, GPIO.LOW) GPIO.output(24, GPIO.HIGH) def RIGHT(): GPIO.output(23, GPIO.HIGH) GPIO.output(24, GPIO.LOW) def STOP(): GPIO.output(23, GPIO.HIGH) GPIO.output(24, GPIO.HIGH) def track(): while close==0: gray = cv2.cvtColor(faceImg,cv2.COLOR_BGR2GRAY) faceRects = classifier.detectMultiScale(gray,scaleFactor=1.2,minNeighbors=3,minSize=(32, 32)) if len(faceRects): x,y,w,h = faceRects[0] # 框选出人脸 最后一个参数2是框线宽度 # cv2.rectangle(faceImg,(x, y), (x + w, y + h), (0,255,0), 2) central_point = x+w/2 if central_point > left_point: LEFT() print("Left") elif central_point < right_point: RIGHT() print("Right") else: STOP() print("Central") STOP() print("Stop") thread1 = threading.Thread(target=track) thread1.start() # 循环读取图像 while True: faceImg = cap.read()[1] # 读取摄像头图像 cv2.imshow("faceImg",cv2.flip(faceImg,1)) if cv2.waitKey(10) == 27: # 通过esc键退出摄像 break # 关闭摄像头 cap.release() cv2.destroyAllWindows() close=1 STOP() print("Stop")