Python学习笔记（使用百度AI进行是否带口罩识别）

效果：

代码：

#!/usr/bin/python
# -*- coding: UTF-8 -*-
import cv2
import json
from aip import AipBodyAnalysis
import rospy
import sys
reload(sys)
sys.setdefaultencoding( "utf-8" )

def get_image():
	cap = cv2.VideoCapture(1) #打开摄像头
	while(1):
		ret, frame = cap.read() #获取图片
		cv2.imshow('video',frame) #打开窗口显示图片
		if cv2.waitKey(1) & 0xFF == ord('q'): # 当按下q时保存图片并退出显示窗口
			cv2.imwrite("test1.jpg",frame)
			break
	cap.release()
	cv2.destroyAllWindows()

""" 读取图片 """
def get_file_content(filePath):
	with open(filePath, 'rb') as fp:
		return fp.read()

# 检测图片
def chack_image():
	
	image = get_file_content('test1.jpg')

	#ID setup
	APP_ID = '19103828'
	API_KEY = 'vxYkTAqGYoWe4XRonMUeCvtH'
	SECRET_KEY = 'etnQGZksMjPFU7GcGUIwtVOL7MBUK0Pa'

	client = AipBodyAnalysis(APP_ID, API_KEY, SECRET_KEY)
	
	""" 调用人体关键点识别 """
	keyword = client.bodyAttr(image)

	""" 打印json格式 """
	print json.dumps(keyword, sort_keys = True, indent = 2)

	print("person_num:%d" %(keyword['person_num']))

	draw_image = cv2.imread('test1.jpg')

	for i in range(0, keyword['person_num']):
		#识别人体的位置（x，y）
		x1 = keyword['person_info'][i]['location']['left']
		y1 = keyword['person_info'][i]['location']['top']

		x2 = keyword['person_info'][i]['location']['width'] + x1
		y2 = keyword['person_info'][i]['location']['height'] + y1

		
		#点位取整
		round(x1)
		round(x2)
		round(y1)
		round(y2)
	
        	#打开图片并绘画出来人体框图
		print("draw_image")

		print("( %d , %d )" %(x1,y1))
		print("( %d , %d )" %(x2,y2))	
	
		#画出矩形
		cv2.rectangle(draw_image, (int(x1),int(y1)), (int(x2),int(y2)), (0,255,0), 4)
		
		face_mask = keyword['person_info'][i]['attributes']['face_mask']['name']
		face_mask = face_mask.decode('utf-8')
		mask1 = '无口罩'
		mask1.decode('utf-8')
		mask2 = '戴口罩'
		mask2.decode('utf-8')
		mask3 = '不确定'
		mask3.decode('utf-8')

		if face_mask == mask1:
			mask = "No mask"
		if face_mask == mask2:
			mask = "wear mask"
		if face_mask == mask3:
			mask = "uncertain"


		print(face_mask)
		person_name = 'person'+str(i) + mask
		
		#标注文本
		cv2.putText(draw_image, person_name, (int(x1),int(y1)), cv2.FONT_HERSHEY_SIMPLEX, 1, (0,0,255), 3)

	while(1):
		cv2.imshow('porson_detect',draw_image) #打开窗口显示图片
		if cv2.waitKey(1) & 0xFF == ord('w'): # 当按下w时保存图片并退出显示窗口
			cv2.imwrite("test1.jpg",draw_image)
			break
	cv2.destroyAllWindows()

if __name__ == '__main__':
	get_image()
	chack_image()