目录：
1.打开摄像头
import cv2 dispW=640 dispH=480 flip=4 #CSI摄像头 camSet='nvarguscamerasrc ! video/x-raw(memory:NVMM), width=3264, height=2464, format=NV12, framerate=21/1 ! nvvidconv flip-method='+str(flip)+' ! video/x-raw, width='+str(dispW)+', height='+str(dispH)+', format=BGRx ! videoconvert ! video/x-raw, format=BGR ! appsink' cam = cv2.VideoCapture(camSet) while True: ret, frame = cam.read() cv2.imshow('frame', frame) if cv2.waitKey(5) & 0xFF == 27: break cam.release() cv2.destroyAllWindows()
import cv2 # 需关闭mjpg服务 ps-ef|grep mjpg 查看摄像头编号 cat /dev/video cap = cv2.VideoCapture(1) #打开摄像头 最大范围 640×480 cap.set(3,320) #设置画面宽度 cap.set(4,240) #设置画面高度 while 1: #进入无限循环 ret,frame = cap.read() #将摄像头拍摄到的画面作为frame的值，ret为布尔值，如果正确读取帧，返回True #反转视频镜头 #frame = cv2.flip(frame, -1) cv2.imshow('usb frame',frame) #将具体的测试效果显示出来 if cv2.waitKey(5) & 0xFF == 27: #如果按了ESC就退出 当然也可以自己设置 break cap.release() cv2.destroyAllWindows()
2.颜色检测
import cv2 #导入库 import numpy as np import time import threading import os import servo_control from pygame_test import play #关于色域范围可以百度 HSV #百科：https://baike.baidu.com/item/HSV/547122?fr=aladdin #参考：https://blog.csdn.net/leo_888/article/details/88284251 #启动脚本杀死影响程序的进程 os.system('./killmjpg.sh') # 要识别的颜色字典 color_dist = { 'red': {'Lower': np.array([156, 60, 60]), 'Upper': np.array([180, 255, 255])}, 'green': {'Lower': np.array([35, 43, 35]), 'Upper': np.array([90, 255, 255])}, 'blue': {'Lower': np.array([110, 80, 80]), 'Upper': np.array([128, 255, 255])}, # 'gray': {'Lower': np.array([0,0,30]),'Upper':np.array([255,40,80]) } }; color_count = [0,0,0] flag = -1 #flag_bak = -1 running = True #使用之前 关闭mjpg进程 ps -elf|grep mjpg 找到进程号，杀死进程 sudo kill -9 xxx xxx代表进程号 #cat /dev/video img_w = 320 img_h = 240 cap = cv2.VideoCapture(0) #打开摄像头 最大范围 640×480 #cap = cv2.VideoCapture(1) #打开摄像头 最大范围 640×480 cap.set(3,img_w) #设置画面宽度 cap.set(4,img_h) #设置画面高度 def music(flag): global running running = False if flag == 0: play('./audios/0016.wav') print('red') elif flag == 1: play('./audios/0017.wav') print('green') elif flag == 2: play('./audios/0018.wav') print('blue') running = True #eEvent = threading.Event() #th1 = threading.Thread(target=music,args=()) lastTime = time.time(); time.sleep(1) #无限循环 while 1: #进入无线循环 #将摄像头拍摄到的画面作为frame的值 ret,frame = cap.read() if running: for i in color_dist: #1-高斯滤波GaussianBlur() 让图片模糊 frame = cv2.GaussianBlur(frame,(5,5),0) #2-转换HSV的样式 以便检测 hsv = cv2.cvtColor(frame,cv2.COLOR_BGR2HSV) #3-查找字典 mask = cv2.inRange(hsv, color_dist[i]['Lower'], color_dist[i]['Upper']) #4-腐蚀图像 mask = cv2.erode(mask,None,iterations=2) #高斯模糊 mask = cv2.GaussianBlur(mask,(3,3),0) #图像合并 res = cv2.bitwise_and(frame,frame,mask=mask) #6-边缘检测 cnts = cv2.findContours(mask.copy(),cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)[-2] if len(cnts) >0 : #通过边缘检测来确定所识别物体的位置信息得到相对坐标 cnt = max(cnts,key=cv2.contourArea) rect = cv2.minAreaRect(cnt) # 获取最小外接矩形的4个顶点 box = cv2.boxPoints(rect) #获取坐标 长宽 角度 c_x, c_y = rect[0] c_h, c_w = rect[1] c_angle = rect[2] #print(c_x,c_y,c_h,c_w) #判断是否是方块大小且在方框中 if 25 < c_h < 75 and 25 < c_w < 75 and 50 < c_x < 270 and 60 < c_y < 180: #print('---',c_x,c_y,c_h,c_w) #绘制轮廓 cv2.drawContours(frame, [np.int0(box)], -1, (0, 255, 255), 2) #print(time.time(), 'x=', int(c_x), 'y=', int(c_y), 'c_h=', int(c_h), 'c_w=', int(c_w), 'angle=', int(c_angle)) if(time.time() - lastTime > 0.1) : lastTime = time.time() index = -1 if i=='red': index = 0; elif i=='green': index = 1 elif i=='blue': index = 2 # elif i == 'gray': # index = 3 flag = index if index >= 0: color_count[index] += 1 if color_count[index] > 4: color_count[0] = 0 color_count[1] = 0 color_count[2] = 0 th1 = threading.Thread(target=music,args=(flag,)) th1.setDaemon(True) th1.start() else: pass cv2.imshow('frame',frame) #将具体的测试效果显示出来 #cv2.imshow('mask',mask) #cv2.imshow('res',res) if cv2.waitKey(5) & 0xFF == 27: #如果按了ESC就退出 当然也可以自己设置 break cap.release() cv2.destroyAllWindows() #后面两句是常规操作,每次使用摄像头都需要这样设置一波
import cv2 #导入库 import numpy as np import time import threading import os import servo_control from zhinengjiaohu import play #创建显示控件 from jetcam.utils import bgr8_to_jpeg import traitlets import ipywidgets.widgets as widgets from IPython.display import display disp = widgets.Image(format='jpeg', width=320, height=240) display(disp) #关于色域范围可以百度 HSV #百科：https://baike.baidu.com/item/HSV/547122?fr=aladdin #参考：https://blog.csdn.net/leo_888/article/details/88284251 #启动脚本杀死影响程序的进程 os.system('./killmjpg.sh') # 要识别的颜色字典 color_dist = { 'red': {'Lower': np.array([156, 60, 60]), 'Upper': np.array([180, 255, 255])}, 'green': {'Lower': np.array([35, 43, 35]), 'Upper': np.array([90, 255, 255])}, 'blue': {'Lower': np.array([110, 80, 80]), 'Upper': np.array([128, 255, 255])}, # 'gray': {'Lower': np.array([0,0,30]),'Upper':np.array([255,40,80]) } }; color_count = [0,0,0] flag = -1 #flag_bak = -1 running = True #使用之前 关闭mjpg进程 ps -elf|grep mjpg 找到进程号，杀死进程 sudo kill -9 xxx xxx代表进程号 #cat /dev/video img_w = 320 img_h = 240 cap = cv2.VideoCapture(0) #打开摄像头 最大范围 640×480 #cap = cv2.VideoCapture(1) #打开摄像头 最大范围 640×480 cap.set(3,img_w) #设置画面宽度 cap.set(4,img_h) #设置画面高度 def music(flag): global running running = False if flag == 0: play('./audios/0016.wav') print('red') elif flag == 1: play('./audios/0017.wav') print('green') elif flag == 2: play('./audios/0018.wav') print('blue') running = True #eEvent = threading.Event() #th1 = threading.Thread(target=music,args=()) lastTime = time.time(); time.sleep(1) #无限循环 while 1: #进入无线循环 #将摄像头拍摄到的画面作为frame的值 ret,frame = cap.read() if running: for i in color_dist: #1-高斯滤波GaussianBlur() 让图片模糊 frame = cv2.GaussianBlur(frame,(5,5),0) #2-转换HSV的样式 以便检测 hsv = cv2.cvtColor(frame,cv2.COLOR_BGR2HSV) #3-查找字典 mask = cv2.inRange(hsv, color_dist[i]['Lower'], color_dist[i]['Upper']) #4-腐蚀图像 mask = cv2.erode(mask,None,iterations=2) #高斯模糊 mask = cv2.GaussianBlur(mask,(3,3),0) #图像合并 res = cv2.bitwise_and(frame,frame,mask=mask) #6-边缘检测 cnts = cv2.findContours(mask.copy(),cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)[-2] if len(cnts) >0 : #通过边缘检测来确定所识别物体的位置信息得到相对坐标 cnt = max(cnts,key=cv2.contourArea) rect = cv2.minAreaRect(cnt) # 获取最小外接矩形的4个顶点 box = cv2.boxPoints(rect) #获取坐标 长宽 角度 c_x, c_y = rect[0] c_h, c_w = rect[1] c_angle = rect[2] #print(c_x,c_y,c_h,c_w) #判断是否是方块大小且在方框中 if 25 < c_h < 75 and 25 < c_w < 75 and 50 < c_x < 270 and 60 < c_y < 180: #print('---',c_x,c_y,c_h,c_w) #绘制轮廓 cv2.drawContours(frame, [np.int0(box)], -1, (0, 255, 255), 2) #print(time.time(), 'x=', int(c_x), 'y=', int(c_y), 'c_h=', int(c_h), 'c_w=', int(c_w), 'angle=', int(c_angle)) if(time.time() - lastTime > 0.1) : lastTime = time.time() index = -1 if i=='red': index = 0; elif i=='green': index = 1 elif i=='blue': index = 2 # elif i == 'gray': # index = 3 flag = index if index >= 0: color_count[index] += 1 if color_count[index] > 4: color_count[0] = 0 color_count[1] = 0 color_count[2] = 0 th1 = threading.Thread(target=music,args=(flag,)) th1.setDaemon(True) th1.start() else: pass disp.value = bgr8_to_jpeg(frame) #cv2.imshow('frame',frame) #将具体的测试效果显示出来 #if cv2.waitKey(5) & 0xFF == 27: #如果按了ESC就退出 当然也可以自己设置 #break cap.release() #cv2.destroyAllWindows() #后面两句是常规操作,每次使用摄像头都需要这样设置一波
import pygame import time pygame.mixer.init() def play(filename): t1 = time.time() pygame.mixer.music.load(filename) pygame.mixer.music.play() while pygame.mixer.music.get_busy(): #检查是否正在播放音乐 pass #time.sleep(5) t2 = time.time() t = t2 - t1 print('t', t) print('end') if __name__ == '__main__': play('./audios/0018.wav')
3.二维码识别
# 载入必要的库 import os import cv2 import numpy as np import sys import time import servo_control # 启动脚本杀死影响程序的进程 os.system('./killmjpg.sh') time.sleep(0.1) data_old = '' data_new = '' cam = cv2.VideoCapture(0) #cam = cv2.VideoCapture(1) # 摄像头二维码实时识别 def Video_display(): global data_old,data_new qrCodeDetector = cv2.QRCodeDetector() while 1: ret, frame = cam.read() gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) t = time.time() data, points, _ = qrCodeDetector.detectAndDecode(gray) #print("Time Taken for Detect and Decode : {:.3f} seconds".format(time.time() - t)) if points is not None: #print('p', tuple(points[0][0])) nrOfPoints = len(points) for i in range(nrOfPoints): nextPointIndex = (i+1) % nrOfPoints cv2.line(frame, tuple(points[i][0]), tuple(points[nextPointIndex][0]), (255,0,0), 5) if data: data_new = data if data_old != data_new: print('data: ', data) data_old = data_new cv2.putText(frame, data, tuple(points[0][0]), cv2.FONT_HERSHEY_SIMPLEX, 1, (0,0,0), 2) cv2.imshow('frame: ', frame) if cv2.waitKey(5) & 0xFF == 27: break if __name__ == '__main__': try: Video_display() except KeyboardInterrupt: cam.release() cv2.destroyAllWindows()
# 载入必要的库 import os import cv2 import numpy as np import sys import time import servo_control #创建显示控件 from jetcam.utils import bgr8_to_jpeg import traitlets import ipywidgets.widgets as widgets from IPython.display import display disp = widgets.Image(format='jpeg', width=320, height=240) display(disp) # 启动脚本杀死影响程序的进程 os.system('./killmjpg.sh') time.sleep(0.1) data_old = '' data_new = '' cam = cv2.VideoCapture(0) #cam = cv2.VideoCapture(1) # 摄像头二维码实时识别 def Video_display(): global data_old,data_new qrCodeDetector = cv2.QRCodeDetector() while 1: ret, frame = cam.read() gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) t = time.time() data, points, _ = qrCodeDetector.detectAndDecode(gray) #print("Time Taken for Detect and Decode : {:.3f} seconds".format(time.time() - t)) if points is not None: #print('p', tuple(points[0][0])) nrOfPoints = len(points) for i in range(nrOfPoints): nextPointIndex = (i+1) % nrOfPoints cv2.line(frame, tuple(points[i][0]), tuple(points[nextPointIndex][0]), (255,0,0), 7) if data: data_new = data if data_old != data_new: print('data: ', data) data_old = data_new cv2.putText(frame, data, tuple(points[0][0]), cv2.FONT_HERSHEY_SIMPLEX, 1, (0,0,0), 2) #cv2.imshow('frame: ', frame) disp.value = bgr8_to_jpeg(frame) #if cv2.waitKey(5) & 0xFF == 27: #break if __name__ == '__main__': try: Video_display() except KeyboardInterrupt: cam.release()
# 载入必要的库 import cv2 import numpy as np import sys import time from baiduaip_test import text2audio,voice import threading text = '' text_old = '' cam = cv2.VideoCapture(0) def music(text): global text_old if text_old != text: text_old = text print('data', data) filepath = text2audio('二维码内容为' + text) voice(filepath) # 摄像头二维码实时识别 def Video_display(): qrCodeDetector = cv2.QRCodeDetector() global text, text_old while 1: ret, frame = cam.read() gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) #t = time.time() data, points, _ = qrCodeDetector.detectAndDecode(gray) #print("Time Taken for Detect and Decode : {:.3f} seconds".format(time.time() - t)) if points is not None: #print('p', tuple(points[0][0])) nrOfPoints = len(points) for i in range(nrOfPoints): nextPointIndex = (i+1) % nrOfPoints cv2.line(frame, tuple(points[i][0]), tuple(points[nextPointIndex][0]), (255,0,0), 5) if data: cv2.putText(frame, data, tuple(points[0][0]), cv2.FONT_HERSHEY_SIMPLEX, 1, (0,0,0), 2) t = threading.Thread(target=music, args=(data,)) t.setDaemon(True) t.start() cv2.imshow('frame: ', frame) if cv2.waitKey(5) & 0xFF == 27: break if __name__ == '__main__': try: Video_display() except KeyboardInterrupt: cam.release() cv2.destroyAllWindows()
import os import time import pygame from aip import AipSpeech import pyaudio_test """ 你的 APPID AK SK """ APP_ID = '24705618' API_KEY = 'D9nixLbfFdTp4B7rT378Y67K' SECRET_KEY = '2Rm19eciqEa9sCUHNyls50QUNtU34qeu' client = AipSpeech(APP_ID, API_KEY, SECRET_KEY) # 读取文件 def get_file_content(filePath): os.system("ffmpeg -y -i {} -acodec pcm_s16le -f s16le -ac 1 -ar 16000 {}.pcm".format(filePath,filePath)) with open("{}.pcm".format(filePath), 'rb') as fp: return fp.read() # 语音识别 def audio2text(filepath): # 识别本地文件 res = client.asr(get_file_content(filepath), 'pcm', 16000, {'dev_pid': 1537,}) print('res', res) #text = res.get('result'][0] if not res.get('err_no') else 'error' return res.get("result")[0] if res.get('result') else 'error' # 语音合成 def text2audio(text): #filename = "{}.mp3".format(int(time.time())) filename = '1.mp3' path = '/tmp/' filepath = os.path.join(path, filename) result = client.synthesis(text, 'zh', 1, {'vol': 7,"spd": 4,"pit": 7,"per": 4}) # 识别正确返回语音二进制 错误则返回dict 参照下面错误码 if not isinstance(result, dict): #print('start write') with open(filepath, 'wb') as f: f.write(result) return filepath # 语音播放 def voice(filepath): print('voice start') pygame.mixer.init() pygame.mixer.music.load(filepath) pygame.mixer.music.play() while pygame.mixer.music.get_busy(): #检查是否正在播放音乐 pass print('voice end') if __name__ == '__main__': filepath = pyaudio_test.rec('1.wav') text = audio2text(filepath) #print('text', text) filename = text2audio(text) voice(filename)
import cv2 import numpy as np import pyzbar.pyzbar as pyzbar import time ''' #调用csi摄像头 dispW=320 dispH=240 flip=4 camSet='nvarguscamerasrc ! video/x-raw(memory:NVMM), width=3264, height=2464, format=NV12, framerate=21/1 ! nvvidconv flip-method='+str(flip)+' ! video/x-raw, width='+str(dispW)+', height='+str(dispH)+', format=BGRx ! videoconvert ! video/x-raw, format=BGR ! appsink' cam = cv2.VideoCapture(camSet) ''' # 调用USB摄像头 cam = cv2.VideoCapture(1) width = cam.get(cv2.CAP_PROP_FRAME_WIDTH) height = cam.get(cv2.CAP_PROP_FRAME_HEIGHT) print('width:', width, 'height:', height) while True: # 读取当前帧 ret, frame = cam.read() # 转为灰度图像 gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) barcodes = pyzbar.decode(gray) #解析图片信息 for barcode in barcodes: # 提取条形码的边界框的位置 # 画出图像中条形码的边界框 (x, y, w, h) = barcode.rect cv2.rectangle(gray, (x, y), (x + w, y + h), (0, 0, 255), 2) # 条形码数据为字节对象，所以如果我们想在输出图像上 # 画出来，就需要先将它转换成字符串 barcodeData = barcode.data.decode("utf-8") barcodeType = barcode.type # 绘出图像上条形码的数据和条形码类型 text = "{} ({})".format(barcodeData, barcodeType) cv2.putText(gray, text, (x, y - 10), cv2.FONT_HERSHEY_SIMPLEX, .5, (0, 0, 125), 2) # 向终端打印条形码数据和条形码类型 print(time.time(), "[INFO] Found {} barcode: {}".format(barcodeType, barcodeData)) cv2.imshow("camera", gray) if cv2.waitKey(5) & 0xFF == 27: #如果按了ESC就退出 当然也可以自己设置 break #关闭摄像头 camera.release() #关闭窗口 cv2.destroyAllWindows()
4.人脸识别
# 导入必要的库 import os import numpy as np import cv2 import servo_control # 载入HAAR模型 #face_cascade = cv2.CascadeClassifier('./trains/haarcascade_frontalface_default.xml') #eye_cascade = cv2.CascadeClassifier('./trains/haarcascade_eye.xml') # 启动脚本杀死影响程序的进程 os.system('./killmjpg.sh') face_cascade = cv2.CascadeClassifier('./trains/face.xml') eye_cascade = cv2.CascadeClassifier('./trains/eye.xml') # csi camera ''' dispW=480 dispH=320 flip=4 #Uncomment These next Two Line for Pi Camera camSet='nvarguscamerasrc ! video/x-raw(memory:NVMM), width=3264, height=2464, format=NV12, framerate=21/1 ! nvvidconv flip-method='+str(flip)+' ! video/x-raw, width='+str(dispW)+', height='+str(dispH)+', format=BGRx ! videoconvert ! video/x-raw, format=BGR ! appsink' cap= cv2.VideoCapture(camSet) ''' # usb camera cap = cv2.VideoCapture(0) #cap = cv2.VideoCapture(1) dis_w = 320 dis_h = 240 cap.set(3, dis_w) cap.set(4, dis_h) def Video_display(): while 1: ret, img = cap.read() img = cv2.flip(img,1) gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) faces = face_cascade.detectMultiScale(gray, 1.3, 5) for (x,y,w,h) in faces: cv2.rectangle(img,(x,y),(x+w,y+h),(255,0,0),2) roi_gray = gray[y:y+h, x:x+w] roi_color = img[y:y+h, x:x+w] print(int(x+w/2), int(y+h/2)) eyes = eye_cascade.detectMultiScale(roi_gray) for (ex,ey,ew,eh) in eyes: cv2.rectangle(roi_color,(ex,ey),(ex+ew,ey+eh),(0,255,0),2) cv2.imshow('img ', img) if cv2.waitKey(5) & 0xFF ==27: break if __name__ == '__main__': try: Video_display() except KeyboardInterrupt: cap.release() cv2.destroyAllWindows()
# 导入必要的库 import os import numpy as np import cv2 import servo_control # 载入HAAR模型 #face_cascade = cv2.CascadeClassifier('./trains/haarcascade_frontalface_default.xml') #eye_cascade = cv2.CascadeClassifier('./trains/haarcascade_eye.xml') #创建显示控件 from jetcam.utils import bgr8_to_jpeg import traitlets import ipywidgets.widgets as widgets from IPython.display import display disp = widgets.Image(format='jpeg', width=320, height=240) display(disp) # 启动脚本杀死影响程序的进程 os.system('./killmjpg.sh') face_cascade = cv2.CascadeClassifier('./trains/face.xml') eye_cascade = cv2.CascadeClassifier('./trains/eye.xml') # csi camera ''' dispW=480 dispH=320 flip=4 #Uncomment These next Two Line for Pi Camera camSet='nvarguscamerasrc ! video/x-raw(memory:NVMM), width=3264, height=2464, format=NV12, framerate=21/1 ! nvvidconv flip-method='+str(flip)+' ! video/x-raw, width='+str(dispW)+', height='+str(dispH)+', format=BGRx ! videoconvert ! video/x-raw, format=BGR ! appsink' cap= cv2.VideoCapture(camSet) ''' # usb camera cap = cv2.VideoCapture(0) #cap = cv2.VideoCapture(1) dis_w = 320 dis_h = 240 cap.set(3, dis_w) cap.set(4, dis_h) def Video_display(): while 1: ret, img = cap.read() img = cv2.flip(img,1) gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) faces = face_cascade.detectMultiScale(gray, 1.3, 5) for (x,y,w,h) in faces: cv2.rectangle(img,(x,y),(x+w,y+h),(255,0,0),2) roi_gray = gray[y:y+h, x:x+w] roi_color = img[y:y+h, x:x+w] print(int(x+w/2), int(y+h/2)) eyes = eye_cascade.detectMultiScale(roi_gray) for (ex,ey,ew,eh) in eyes: cv2.rectangle(roi_color,(ex,ey),(ex+ew,ey+eh),(0,255,0),2) #cv2.imshow('img ', img) disp.value = bgr8_to_jpeg(img) #if cv2.waitKey(5) & 0xFF ==27: #break if __name__ == '__main__': try: Video_display() except KeyboardInterrupt: cap.release()
5.行人检测
import cv2 import numpy as np # 创建我们的身体分类器 body_classifier = cv2.CascadeClassifier('./images/haarcascade_fullbody.xml') # 为视频文件启动视频捕获 cap = cv2.VideoCapture('./images/walking.avi') #一旦视频成功加载，循环播放 def Video_display(): while cap.isOpened(): # Read first frame ret, frame = cap.read() frame = cv2.resize(frame, None,fx=0.5, fy=0.5, interpolation = cv2.INTER_LINEAR) gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) # Pass frame to our body classifier bodies = body_classifier.detectMultiScale(gray, 1.2, 3) # Extract bounding boxes for any bodies identified for (x,y,w,h) in bodies: cv2.rectangle(frame, (x, y), (x+w, y+h), (0, 255, 255), 2) cv2.imshow('frame', frame) if cv2.waitKey(5) & 0xFF == 27: break cap.release() cv2.destroyAllWindows() if __name__ == '__main__': Video_display()
6.汽车检测
# 载入必要的库 import os import cv2 import time import numpy as np import servo_control # 启动脚本杀死影响程序的进程 os.system('./killmjpg.sh') time.sleep(0.1) # 载入HAAR分类器 car_classifier = cv2.CascadeClassifier('./trains/haarcascade_car.xml') cap = cv2.VideoCapture(0) #cap = cv2.VideoCapture(1) def car_detect(): while 1: time.sleep(.05) # Read first frame ret, frame = cap.read() gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) # Pass frame to our car classifier cars = car_classifier.detectMultiScale(gray, 1.4, 2) # Extract bounding boxes for any bodies identified for (x,y,w,h) in cars: cv2.rectangle(frame, (x, y), (x+w, y+h), (0, 255, 255), 2) cv2.imshow('frame', frame) if cv2.waitKey(5) & 0xFF == 27: break if __name__ == '__main__': try: car_detect() except KeyboardInterrupt: cap.release() cv2.destroyAllWindows()
# 载入必要的库 import os import cv2 import time import numpy as np import servo_control #创建显示控件 from jetcam.utils import bgr8_to_jpeg import traitlets import ipywidgets.widgets as widgets from IPython.display import display disp = widgets.Image(format='jpeg', width=320, height=240) display(disp) # 启动脚本杀死影响程序的进程 os.system('./killmjpg.sh') time.sleep(0.1) # 载入HAAR分类器 car_classifier = cv2.CascadeClassifier('./trains/haarcascade_car.xml') cap = cv2.VideoCapture(0) #cap = cv2.VideoCapture(1) def car_detect(): try: while 1: time.sleep(.05) # Read first frame ret, frame = cap.read() gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) # Pass frame to our car classifier cars = car_classifier.detectMultiScale(gray, 1.4, 2) # Extract bounding boxes for any bodies identified for (x,y,w,h) in cars: cv2.rectangle(frame, (x, y), (x+w, y+h), (0, 255, 255), 2) disp.value = bgr8_to_jpeg(frame) #cv2.imshow('frame', frame) #if cv2.waitKey(5) & 0xFF == 27: #break except KeyboardInterrupt: cap.release() car_detect()
7.车牌检测
import cv2 import imutils import numpy as np import pytesseract img = cv2.imread('./images/makecar.jpg',cv2.IMREAD_COLOR) img = cv2.resize(img, (600,400) ) gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) gray = cv2.bilateralFilter(gray, 13, 15, 15) edged = cv2.Canny(gray, 30, 200) contours = cv2.findContours(edged.copy(), cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE) contours = imutils.grab_contours(contours) contours = sorted(contours, key = cv2.contourArea, reverse = True)[:10] screenCnt = None for c in contours: peri = cv2.arcLength(c, True) approx = cv2.approxPolyDP(c, 0.018 * peri, True) if len(approx) == 4: screenCnt = approx break if screenCnt is None: detected = 0 print ("No contour detected") else: detected = 1 if detected == 1: cv2.drawContours(img, [screenCnt], -1, (0, 0, 255), 3) mask = np.zeros(gray.shape,np.uint8) new_image = cv2.drawContours(mask,[screenCnt],0,255,-1,) new_image = cv2.bitwise_and(img,img,mask=mask) (x, y) = np.where(mask == 255) (topx, topy) = (np.min(x), np.min(y)) (bottomx, bottomy) = (np.max(x), np.max(y)) Cropped = gray[topx:bottomx+1, topy:bottomy+1] text = pytesseract.image_to_string(Cropped, config='--psm 11') print("programming_fever's License Plate Recognition\n") print("Detected license plate Number is:",text) img = cv2.resize(img,(500,300)) Cropped = cv2.resize(Cropped,(400,200)) # 显示图像 try: cv2.imshow('img', img) cv2.imshow('cropped', Cropped) # if cv2.waitKey(5) & 0xFF ==27: # break finally: cv2.destroyAllWindows()
8.OpenCV根据颜色定位物体的实时位置
from __future__ import print_function from imutils.video import VideoStream import imutils import time import cv2 import os import RPi.GPIO as GPIO # LED 初始化 redLed = 17 GPIO.setwarnings(False) GPIO.setmode(GPIO.BCM) GPIO.setup(redLed, GPIO.OUT) # 打印物体的实时位置 def mapObjectPosition (x, y): print ("[INFO] Object Center coordenates at X0 = {0} and Y0 = {1}".format(x, y)) ''' dispW=500 dispH=350 flip=4 # 调用CSI摄像头的属性参数 camSet=camSet='nvarguscamerasrc ! video/x-raw(memory:NVMM), width=3264, height=2464, format=NV12, framerate=21/1 ! nvvidconv flip-method='+str(flip)+' ! video/x-raw, width='+str(dispW)+', height='+str(dispH)+', format=BGRx ! videoconvert ! video/x-raw, format=BGR ! appsink' # 初始化视频流并允许摄像机传感器预热 print("[INFO] waiting for camera to warmup...") vs = cv2.VideoCapture(camSet) #VideoStream(camSet).start() ''' vs = cv2.VideoCapture(1) time.sleep(2.0) # 定义对象的上下边界 # 在HSV颜色空间中进行跟踪 colorLower = (9,135,231) colorUpper = (31,255,255) # 关闭LED灯 GPIO.output(redLed, GPIO.LOW) ledOn = False def Video_display(): global ledOn # 循环的帧从视频流 while True: # 从视频流中抓取下一帧，调整大小 # 帧，并将其转换为HSV颜色空间 ret, frame = vs.read() frame = imutils.resize(frame, width=500) frame = imutils.rotate(frame, angle=0) hsv = cv2.cvtColor(frame, cv2.COLOR_BGR2HSV) # 为对象颜色构造一个遮罩，然后执行 # 一系列的膨胀和侵蚀，以消除任何小的 # blobs left in the mask mask = cv2.inRange(hsv, colorLower, colorUpper) mask = cv2.erode(mask, None, iterations=2) mask = cv2.dilate(mask, None, iterations=2) # 找到遮罩中的轮廓并初始化 # (x, y) center of the object cnts = cv2.findContours(mask.copy(), cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)[-2] #cnts = cnts[0] if imutils.is_cv2() else cnts[1] center = None # 只有在找到至少一条轮廓线时才进行 if len(cnts) > 0: # 在蒙版中找到最大的轮廓，然后使用 # 它可以计算出最小的围圆 # 重心 c = max(cnts, key=cv2.contourArea) ((x, y), radius) = cv2.minEnclosingCircle(c) M = cv2.moments(c) center = (int(M["m10"] / M["m00"]), int(M["m01"] / M["m00"])) # 只有当半径满足最小尺寸时才进行 if radius > 10: # 在框架上画圆和质心， # 然后更新跟踪点的列表 cv2.circle(frame, (int(x), int(y)), int(radius),(0, 255, 255), 2) cv2.circle(frame, center, 5, (0, 0, 255), -1) # 定位舵机在圆心 mapObjectPosition(int(x), int(y)) # 如果led还没有打开，打开led if not ledOn: GPIO.output(redLed, GPIO.HIGH) ledOn = True # 如果没有检测到球，关闭LED灯 elif ledOn: GPIO.output(redLed, GPIO.LOW) ledOn = False # 向我们的屏幕显示框架 cv2.imshow('frame', frame) if cv2.waitKey(5) & 0xFF == 27: break # 做点清理工作 print("\n [INFO] Exiting Program and cleanup stuff \n") GPIO.cleanup() vs.stop() vs.release() cv2.destroyAllWindows() if __name__ == '__main__': Video_display()
9.OpenCV单线循迹
''' 图像识别功能使用前，如果树梅派在开机前有连接的摄像头， 需要先关闭mjpg进程 请在光照环境稳定且背景为纯色的情况下使用图像识别功能。 ''' import cv2 #导入库 import numpy as np import time import z_uart as myUart cap = cv2.VideoCapture(0) #打开摄像头 cap.set(3,160) cap.set(4,120) #设置窗口的大小 dis_w = 160 #全局变量 #systick_ms_bak = 0 myUart.setup_uart(115200) def Tracing(cx): print('cx', cx) if cx <= 40: #偏you l_pwm = 1500 + int(600*(cx-40)/40) car_run(-10, 200, l_pwm) elif 40 < cx < 120: car_run(200, 200) elif cx >= 120: #偏zuo r_pwm = 1500 + int(600*(cx-120)/40) car_run(200, -10, r_pwm) def car_run(left_speed, right_speed, pwm_value=1500): testStr = "{#006P%04dT0000!#007P%04dT0000!#000P%04dT0000!}" % (1500+left_speed, 1500-right_speed, pwm_value) print(testStr) myUart.uart_send_str(testStr) def camera_xunji(): while True: #进入无线循环 # Capture the frames ret, frame = cap.read() #ret, img = cap.read()中，cap.read()是按帧读取，其会返回两个值：ret,img（ret是布尔值，如果读取帧是正确的则返回True，如果文件读取到结尾，它的返回值就为False；后面的frame该帧图像的三维矩阵BGR形式。) #反转视频镜头 #frame = cv2.flip(frame, -1) t1 = time.time() # Crop the image crop_img =frame[60:120, 0:160] #frame是相对于父视图的坐标系来定位的。如果你这样设置frame:(0,0,100,200),也就是在父视图左上角添加了一个宽100，高200的子视图（前提是没有改变父视图的bounds，接下来会有介绍bounds）。 #crop_img =frame(0,0,640,480) # 转化成GRAY样式 gray = cv2.cvtColor(crop_img, cv2.COLOR_BGR2GRAY) #高斯滤波GaussianBlur() 让图片模糊 blur = cv2.GaussianBlur(gray,(5,5),0) #二值化 ret,thresh1 = cv2.threshold(blur,60,255,cv2.THRESH_BINARY_INV) #腐蚀，去除图像毛次，减少细微影响 mask = cv2.erode(thresh1, None, iterations=2) #膨胀， mask = cv2.dilate(mask, None, iterations=2) # 查找轮廓 contours = cv2.findContours(mask.copy(), 1, cv2.CHAIN_APPROX_NONE)[-2] # Find the biggest contour (if detected) if len(contours) > 0: c = max(contours, key=cv2.contourArea) M = cv2.moments(c) cx = int(M['m10']/M['m00']) cy = int(M['m01']/M['m00']) #根据得到的坐标画出蓝线，方便看出偏差 cv2.line(crop_img,(cx,0),(cx,720),(255,0,0),1) cv2.line(crop_img,(0,cy),(1280,cy),(255,0,0),1) #绘制绿色轮廓 cv2.drawContours(crop_img, contours, -1, (0,255,0), 1) #print(time.time(), int(cx),int(cy Tracing(cx) else: print("I don't see the line" ) cv2.imshow('frame', crop_img) #将具体的测试效果显示出来 t2=time.time() pfs = 1.0/(t2-t1) print('fps',pfs) #cv2.imshow('mask',mask) #cv2.imshow('res',res) if cv2.waitKey(5) & 0xFF == 27: #如果按了ESC就退出 当然也可以自己设置 break if __name__ == '__main__': try: camera_xunji() except KeyboardInterrupt: car_run(0,0,) cap.release() cv2.destroyAllWindows() #后面两句是常规操作,每次使用摄像头都需要这样设置一波
#导包 import serial import time import threading import z_led as myLed import z_beep as myBeep #全局变量定义 ser = '' uart_baud = 115200 uart_get_ok = 0 uart_receive_buf = "" uart_receive_buf_index = 0 #发送字符串 只需传入要发送的字符串即可 def uart_send_str(string): global ser ser.write(string.encode("GB2312")) #time.sleep(0.01) #ser.flushInput() #线程调用函数，主要处理数据接受格式，主要格式为 $...! #...! {...}三种格式，...内容长度不限 def serialEvent(): global ser,uart_receive_buf_index,uart_receive_buf,uart_get_ok mode = 0 try: while True: if uart_get_ok == 0: uart_receive_buf_index = ser.inWaiting() if uart_receive_buf_index > 0: uart_receive_buf = uart_receive_buf+ser.read(uart_receive_buf_index).decode() #print('get1:',uart_receive_buf, " len:", len(uart_receive_buf), " mode:", mode) if mode == 0: if uart_receive_buf.find('{') >= 0: mode = 1 #print('mode1 start') elif uart_receive_buf.find('$') >= 0: mode = 2 #print('mode2 start') elif uart_receive_buf.find('#') >= 0: mode = 3 #print('mode3 start') if mode == 1: if uart_receive_buf.find('}') >= 0: uart_get_ok = 1 mode = 0 ser.flushInput() #print('{}:',uart_receive_buf, " len:", len(uart_receive_buf)) #print('mode1 end') elif mode == 2: if uart_receive_buf.find('!') >= 0: uart_get_ok = 2 mode = 0 ser.flushInput() #print('$!:',uart_receive_buf, " len:", len(uart_receive_buf)) #print('mode2 end') elif mode == 3: if uart_receive_buf.find('!') >= 0: uart_get_ok = 3 mode = 0 ser.flushInput() #print('#!:', uart_receive_buf, " len:", len(uart_receive_buf)) #print('mode3 end') #print('get2:',uart_receive_buf, " len:", len(uart_receive_buf), " mode:", mode, " getok:", uart_get_ok) except IOError: pass #串口接收线程 uart_thread = threading.Thread(target=serialEvent) #串口初始化 def setup_uart(baud): global ser,uart_thread,uart_receive_buf uart_baud = baud ser = serial.Serial("/dev/ttyTHS1", uart_baud, timeout=2) ser.flushInput() uart_thread.start() uart_send_str("uart init ok!\r\n"); uart_receive_buf = '' #循环执行串口 def loop_uart(): global uart_get_ok, uart_receive_buf if(uart_get_ok): print(int(time.time()*1000)) uart_send_str(uart_receive_buf) if uart_receive_buf == '$LEDON!': myLed.on(1) elif uart_receive_buf == '$LEDOFF!': myLed.off(1) elif uart_receive_buf == '$BEEPON!': myBeep.on() elif uart_receive_buf == '$BEEPOFF!': myBeep.off() uart_receive_buf = '' uart_get_ok = 0 #大循环 if __name__ == '__main__': setup_uart(115200) myLed.setup_led() myBeep.setup_beep() #发出哔哔哔作为开机声音 myBeep.beep(0.1) myBeep.beep(0.1) myBeep.beep(0.1) try: while True: loop_uart() except KeyboardInterrupt: if ser != None: ser.close() myLed.off(1) myLed.off(2) myBeep.off()
10.超声波自由避障
#导入模块 import time import RPi.GPIO as GPIO import z_uart as myUart #端口模式设置 GPIO.setmode(GPIO.BCM) GPIO.setwarnings(False) #引脚定义 BEEP_PIN = 21 LED_PIN = 4 KEY1_PIN = 24 KEY2_PIN = 25 #超声波引脚定义 TRIG = 23 ECHO = 22 #全局变量定义 systick_ms_bak = 0 dis = 100 #初始化IO口 def setup_gpio(): GPIO.setup(LED_PIN, GPIO.OUT,initial=1) GPIO.setup(BEEP_PIN, GPIO.OUT, initial=0) GPIO.setup(KEY1_PIN, GPIO.IN, pull_up_down = GPIO.PUD_UP) GPIO.setup(KEY2_PIN, GPIO.IN, pull_up_down = GPIO.PUD_UP) #初始化超声波 def setup_csb(): GPIO.setup(TRIG, GPIO.OUT,initial = 0) GPIO.setup(ECHO, GPIO.IN,pull_up_down = GPIO.PUD_UP) #大循环LED灯 def loop_led(): global systick_ms_bak if(int((time.time() * 1000))- systick_ms_bak > 500): systick_ms_bak = int((time.time() * 1000)) GPIO.output(LED_PIN, not GPIO.input(LED_PIN)) #led引脚电平翻转 #间接发出响声，x代表间接的时间，单位为秒 def beep(x): GPIO.output(BEEP_PIN,1) #蜂鸣器鸣叫 time.sleep(x) GPIO.output(BEEP_PIN,0) #蜂鸣器关闭 time.sleep(x) #启动示意 def setup_start(): for i in range(3): beep(0.1) #超声波测距 def distance(): GPIO.output(TRIG, 0) time.sleep(0.000002) GPIO.output(TRIG, 1) time.sleep(0.00001) GPIO.output(TRIG, 0) while GPIO.input(ECHO) == 0: a = 0 time1 = time.time() while GPIO.input(ECHO) == 1: a = 1 time2 = time.time() during = time2 - time1 return during * 340 / 2 * 100 #自由避障 def car_move(): global systick_ms_bak if(int((time.time() * 1000))- systick_ms_bak > 100): systick_ms_bak = int((time.time() * 1000)) dis = distance() if int(dis) >= 100: car_go_back(400) elif 50 <= int(dis) < 100: car_go_back(300) elif 20 <= int(dis) < 50: car_go_back(200) else: car_right_turn(500) ''' 函数功能：串口发送指令控制电机转动 范围：-1000～+1000 ''' def car_run(left_speed, right_speed, pwm_value=1500): textSrt = '#006P{:0>4d}T0000!#007P{:0>4d}T0000!#000P{:0>4d}T0000'.format(1500+left_speed, 1500-right_speed, pwm_value) print(textSrt) myUart.uart_send_str(textSrt) ''' 函数功能：小车前进后退 正值小车前进，负值小车后退 范围：-1000～+1000 ''' def car_go_back(speed, pwm_value=1500): car_run(speed, speed) ''' 函数功能：小车右转 正值小车右转 范围：0~1000 ''' def car_right_turn(speed, pwm_value=1500): speedl = speed speedr = speed*2//3 car_run(speedl, speedr, 2100) ''' 函数功能：小车停止 ''' def car_stop(): myUart.uart_send_str('#006P1500T1000!#007P1500T1000!#000P1500T1000!') #关闭 def destory(): GPIO.output(LED_PIN, 1) GPIO.output(BEEP_PIN, 0) GPIO.cleanup() car_stop() #大循环 if __name__ == '__main__': setup_gpio() #初始化IO口 setup_start() #初始化示意蜂鸣器滴滴滴三声 setup_csb() #初始化超声波 myUart.setup_uart(115200) #设置串口 try: while True: loop_led() car_move() except KeyboardInterrupt: destory()
11.智能识物实时播放
# 实时播放识别物体对象 import cv2 import jetson.inference import jetson.utils import time import numpy as np import threading # 线程 from baiduaip_test import voice speak = True item = '欢迎使用众灵AI智能语音识别系统！！' confidence = 0 itemOld='' ''' import pygame from aip import AipSpeech #这里需要填你自己的id和密钥 APP_ID='24705618' API_KEY='D9nixLbfFdTp4B7rT378Y67K' SECRET_KEY='2Rm19eciqEa9sCUHNyls50QUNtU34qeu' aipSpeech=AipSpeech(APP_ID,API_KEY,SECRET_KEY) ''' def sayItem(): global speak global item while True: if speak == True: voice('./audios/zl.mp3') speak=False ''' result = aipSpeech.synthesis(text = item,options={'spd':5,'vol':9,'per':0,}) if not isinstance(result,dict): with open('./audios/zl.mp3','wb') as f: f.write(result) else:print(result) #我们利用Jetson Nano自带的pygame pygame.mixer.init() pygame.mixer.music.load('./audios/zl.mp3') pygame.mixer.music.play() speak=False ''' x = threading.Thread(target=sayItem,daemon=True) x.start() # csi camera ''' dispW=1280 dispH=720 flip=4 # Gstreamer code for improvded Raspberry Pi Camera Quality camSet='nvarguscamerasrc wbmode=3 tnr-mode=2 tnr-strength=1 ee-mode=2 ee-strength=1 ! video/x-raw(memory:NVMM), width=3264, height=2464, format=NV12, framerate=21/1 ! nvvidconv flip-method='+str(flip)+' ! video/x-raw, width='+str(dispW)+', height='+str(dispH)+', format=BGRx ! videoconvert ! video/x-raw, format=BGR ! videobalance contrast=1.5 brightness=-.2 saturation=1.2 ! appsink' cam=cv2.VideoCapture(camSet) ''' # usb camera cam=cv2.VideoCapture(0) cam.set(3,640) cam.set(4,480) dispW = 640 dispH = 480 net=jetson.inference.imageNet('googlenet') font=cv2.FONT_HERSHEY_SIMPLEX timeMark= time.time() fpsFiltered=0 def shiwu(): global speak, item, itemOld, timeMark, fpsFiltered, confidence while True: ret,frame = cam.read() img = cv2.cvtColor(frame,cv2.COLOR_BGR2RGBA).astype(np.float32) img = jetson.utils.cudaFromNumpy(img) if speak == False: classID,confidence = net.Classify(img,dispW,dispH) if confidence>=.5: item=net.GetClassDesc(classID) if item!=itemOld: speak = True if confidence < .5: item='' itemOld = item dt = time.time() - timeMark timeMark = time.time() fps = 1/dt fpsFiltered=.95*fpsFiltered + .05*fps cv2.putText(frame,str(round(fpsFiltered,1)) + ' fps '+item+str(round(confidence,2)),(0,30),font,1,(0,0,255),2) cv2.imshow('frame', frame) if cv2.waitKey(5) & 0xFF == 27: break if __name__ == '__main__': try: shiwu() except KeyboardInterrupt: cam.release() cv2.destroyAllWindows()
import os import time import pygame from aip import AipSpeech import pyaudio_test """ 你的 APPID AK SK """ APP_ID = '24705618' API_KEY = 'D9nixLbfFdTp4B7rT378Y67K' SECRET_KEY = '2Rm19eciqEa9sCUHNyls50QUNtU34qeu' client = AipSpeech(APP_ID, API_KEY, SECRET_KEY) # 读取文件 def get_file_content(filePath): os.system("ffmpeg -y -i {} -acodec pcm_s16le -f s16le -ac 1 -ar 16000 {}.pcm".format(filePath,filePath)) with open("{}.pcm".format(filePath), 'rb') as fp: return fp.read() # 语音识别 def audio2text(filepath): # 识别本地文件 res = client.asr(get_file_content(filepath), 'pcm', 16000, {'dev_pid': 1537,}) print('res', res) #text = res.get('result'][0] if not res.get('err_no') else 'error' return res.get("result")[0] if res.get('result') else 'error' # 语音合成 def text2audio(text): #filename = "{}.mp3".format(int(time.time())) filename = '1.mp3' path = '/tmp/' filepath = os.path.join(path, filename) result = client.synthesis(text, 'zh', 1, {'vol': 7,"spd": 4,"pit": 7,"per": 4}) # 识别正确返回语音二进制 错误则返回dict 参照下面错误码 if not isinstance(result, dict): #print('start write') with open(filepath, 'wb') as f: f.write(result) return filepath # 语音播放 def voice(filepath): print('voice start') pygame.mixer.init() pygame.mixer.music.load(filepath) pygame.mixer.music.play() while pygame.mixer.music.get_busy(): #检查是否正在播放音乐 pass print('voice end') if __name__ == '__main__': filepath = pyaudio_test.rec('1.wav') text = audio2text(filepath) #print('text', text) filename = text2audio(text) voice(filename)
12.人体追踪
# 载入必要的库 import os import time import cv2 import numpy as np import z_uart as myUart import servo_control # 启动脚本杀死影响程序的进程 os.system('./killmjpg.sh') time.sleep(0.1) servo_control.servo_move(0, 90, 1, 60) myUart.setup_uart(115200) # 创建我们的身体分类器 body_classifier = cv2.CascadeClassifier('./trains/haarcascade_fullbody.xml') # 为视频文件启动视频捕获 #cap = cv2.VideoCapture('./videos/walking.avi') cap = cv2.VideoCapture(0) #cap = cv2.VideoCapture(1) dis_w = 640 dis_h = 480 cap.set(3,dis_w) cap.set(4,dis_h) axis_x = dis_w//2//2 def tracing(area, pwm_value): print('area ', area) print('pwm ', pwm_value) if pwm_value < 500: pwm_value = 500 if pwm_value >2500: pwm_value = 2500 if 12000 < area: car_run(-300, -300, pwm_value=pwm_value) elif area <= 12000: car_run(300, 300, pwm_value=pwm_value) def angle(x,y,w,h): print('---',x,y,w,h) center_x = x + w print('cen x ',center_x) if center_x <= axis_x: pwm_value = 1500 - 800*(axis_x-center_x)//axis_x elif axis_x < center_x < axis_x*2: pwm_value = 1500 elif center_x >=axis_x: pwm_value = 1500 - 800*(axis_x - center_x)//axis_x print('pwm_cd', pwm_value) return pwm_value def car_run(left_speed, right_speed, pwm_value=1500): textStr = '{#006P%04dT0000!#007P%04dT0000!#000P%04dT0000!}' % (1500+left_speed, 1500-right_speed, pwm_value) print(textStr) myUart.uart_send_str(textStr) def Video_display(): while 1: # Read first frame ret, frame = cap.read() frame = cv2.resize(frame, None,fx=0.5, fy=0.5, interpolation = cv2.INTER_LINEAR) #print('frame shape ',frame.shape) gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) # Pass frame to our body classifier bodies = body_classifier.detectMultiScale(gray, 1.2, 3) # Extract bounding boxes for any bodies identified for (x,y,w,h) in bodies: print(x,y,w,h) area = int(w*h) if area > 4000: cv2.rectangle(frame, (x, y), (x+w, y+h), (0, 255, 255), 2) pwm_value = angle(x,y,w,h) tracing(area, pwm_value) cv2.imshow('frame', frame) if cv2.waitKey(5) & 0xFF == 27: break if __name__ == '__main__': try: Video_display() except KeyboardInterrupt: car_run(0, 0) cap.release() cv2.destroyAllWindows() finally: car_run(0, 0)
# 载入必要的库 import os import time import cv2 import numpy as np import z_uart as myUart import servo_control #创建显示控件 from jetcam.utils import bgr8_to_jpeg import traitlets import ipywidgets.widgets as widgets from IPython.display import display disp = widgets.Image(format='jpeg', width=640, height=480) display(disp) # 启动脚本杀死影响程序的进程 os.system('./killmjpg.sh') time.sleep(0.1) servo_control.servo_move(0, 90, 1, 60) myUart.setup_uart(115200) # 创建我们的身体分类器 body_classifier = cv2.CascadeClassifier('./trains/haarcascade_fullbody.xml') # 为视频文件启动视频捕获 #cap = cv2.VideoCapture('./videos/walking.avi') cap = cv2.VideoCapture(0) #cap = cv2.VideoCapture(1) dis_w = 640 dis_h = 480 cap.set(3,dis_w) cap.set(4,dis_h) axis_x = dis_w//2//2 def tracing(area, pwm_value): print('area ', area) print('pwm ', pwm_value) if pwm_value < 500: pwm_value = 500 if pwm_value >2500: pwm_value = 2500 if 12000 < area: car_run(-150, -150, pwm_value=pwm_value) elif area <= 12000: car_run(150, 150, pwm_value=pwm_value) def angle(x,y,w,h): print('---',x,y,w,h) center_x = x + w print('cen x ',center_x) if center_x <= axis_x: pwm_value = 1500 - 800*(axis_x-center_x)//axis_x elif axis_x < center_x < axis_x*2: pwm_value = 1500 elif center_x >=axis_x: pwm_value = 1500 - 800*(axis_x - center_x)//axis_x print('pwm_cd', pwm_value) return pwm_value def car_run(left_speed, right_speed, pwm_value=1500): textStr = '{#006P%04dT0000!#007P%04dT0000!#000P%04dT0000!}' % (1500+left_speed, 1500-right_speed, pwm_value) print(textStr) myUart.uart_send_str(textStr) def Video_display(): while 1: # Read first frame ret, frame = cap.read() frame = cv2.resize(frame, None,fx=0.5, fy=0.5, interpolation = cv2.INTER_LINEAR) #print('frame shape ',frame.shape) gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY) # Pass frame to our body classifier bodies = body_classifier.detectMultiScale(gray, 1.2, 3) # Extract bounding boxes for any bodies identified for (x,y,w,h) in bodies: print(x,y,w,h) area = int(w*h) if area > 4000: cv2.rectangle(frame, (x, y), (x+w, y+h), (0, 255, 255), 2) pwm_value = angle(x,y,w,h) tracing(area, pwm_value) disp.value = bgr8_to_jpeg(frame) #cv2.imshow('frame', frame) #if cv2.waitKey(5) & 0xFF == 27: #break if __name__ == '__main__': try: Video_display() except KeyboardInterrupt: car_run(0, 0) cap.release() #cv2.destroyAllWindows() finally: car_run(0, 0)
13.人脸身份识别
基于Dlib玩转人脸身份识别 安装脸部识别库 1.安装依赖库文件 sudo apt-get update sudo apt-get install cmake libopenblas-dev liblapack-dev libjpeg-dev 2.由于nano的内存只有4G，在编译dlib 的过程中，需要一个交换文件 git clone http://github.com/JetsonHacksNano/installSwapfile cd installSwapfile ./installSwapfile.sh 3.下载dlib安装包，解压之后，需要对源代码中一行代码进行注释 wget http://dlib.net/files/dlib-19.17.tar.bz2 tar jxvf dlib-19.17.tar.bz2 cd dlib-19.17 4.修改在dlib/cuda/目录下的cudnn_dlibapi.cpp文件中的内容，使得适用于jetson nano 找到如下目录： /dlib-19.17/dlib/cuda/cudnn_dlibapi.cpp 用jupyterlab打开，删除或注释下面这行代码： forward_algo = forward_best_algo; 5.安装dlib-19.17依赖库 cd .. cd dlib-19.17 sudo python3 setup.py install 执行编译，编译时间较长，请耐心等待 6.安装face_recognition的库 sudo pip3 install face_recognition
# 在OpenCV中训练面部识别模型 # 生成训练模型 import face_recognition import cv2 import os import time import pickle def train(filename='./trains/train.pkl'): Encodings=[] Names=[] j=0 image_dir='./known' for root,dirs,files in os.walk(image_dir): print(files) for file in files: path = os.path.join(root,file) print(path) name=os.path.splitext(file)[0] print(name) person=face_recognition.load_image_file(path) encoding=face_recognition.face_encodings(person)[0] Encodings.append(encoding) Names.append(name) print(Names) with open(filename,'wb') as f: pickle.dump(Names,f) pickle.dump(Encodings,f) return filename cam = cv2.VideoCapture(0) # 通过摄像头进行面部身份识别 def face_detect(filename='./trains/train.pkl'): fpsReport=0 scaleFactor=.25 Encodings=[] Names=[] with open(filename,'rb') as f: Names=pickle.load(f) Encodings=pickle.load(f) timeStamp = time.time() while True: ret,frame = cam.read() frameSmall=cv2.resize(frame,(0,0),fx=scaleFactor,fy=scaleFactor) frameRGB=cv2.cvtColor(frameSmall,cv2.COLOR_BGR2RGB) facePositions=face_recognition.face_locations(frameRGB,model='cnn') allEncodings=face_recognition.face_encodings(frameRGB,facePositions) for (top,right,bottom,left),face_encoding in zip(facePositions,allEncodings): name='Unkown Person' matches=face_recognition.compare_faces(Encodings,face_encoding) if True in matches: first_match_index=matches.index(True) name=Names[first_match_index] top=int(top/scaleFactor) right=int(right/scaleFactor) bottom=int(bottom/scaleFactor) left=int(left/scaleFactor) cv2.rectangle(frame,(left,top),(right,bottom),(0,0,255),2) cv2.putText(frame,name,(left,top-6),cv2.FONT_HERSHEY_SIMPLEX,.75,(0,0,255),2) dt = time.time()-timeStamp fps=1/dt fpsReport=.90*fpsReport + .1*fps #print('fps is:',round(fpsReport)) timeStamp=time.time() cv2.rectangle(frame,(0,0),(100,40),(0,0,255),-1) cv2.putText(frame,str(round(fpsReport,1)) + 'fps',(0,25),cv2.FONT_HERSHEY_SIMPLEX,.75,(0,255,255),2) cv2.imshow('frame', frame) if cv2.waitKey(5) & 0xFF == 27: break if __name__ == '__main__': try: #filename = train() face_detect() except KeyboardInterrupt: cam.release() cv2.destroyAllWindows()
14.识别镜头中感兴趣的对象
import cv2 import jetson.inference import jetson.utils import time import numpy as np timeStamp = time.time() fpsFiltered = 0 obj = 'dog' # interesting thing net = jetson.inference.detectNet('ssd-mobilenet-v2',threshold=.5) font = cv2.FONT_HERSHEY_SIMPLEX cam = cv2.VideoCapture(0) def interest_detect(): global timeStamp, fpsFiltered while True: _,img = cam.read() height = img.shape[0] width = img.shape[1] frame = cv2.cvtColor(img,cv2.COLOR_BGR2RGBA).astype(np.float32) frame = jetson.utils.cudaFromNumpy(frame) detections = net.Detect(frame,width,height) for detect in detections: #print(detect) ID=detect.ClassID top=int(detect.Top) left=int(detect.Left) bottom=int(detect.Bottom) right=int(detect.Right) item=net.GetClassDesc(ID) if item == obj: # 识别interesting thing cv2.rectangle(img,(left,top),(right,bottom),(0,255,0),2) cv2.putText(img,item,(left,top+20),font,1,(0,0,255),2) #print(item,top,left,bottom,right) #print(str(round(fps,1))+' fps ') dt=time.time()-timeStamp timeStamp=time.time() fps=1/dt fpsFiltered=.9*fpsFiltered + .1*fps cv2.putText(img,str(round(fpsFiltered,1)) + ' fps ',(0,30),font,1,(0,0,255),2) cv2.imshow('img', img) if cv2.waitKey(5) & 0xFF == 27: break if __name__ == '__main__': try: interest_detect() except KeyboardInterrupt: cam.release() cv2.destroyAllWindows()
15.颜色跟随
import cv2 #导入库 import numpy as np import time import os from servo_control import * #关于色域范围可以百度 HSV #百科：https://baike.baidu.com/item/HSV/547122?fr=aladdin #参考：https://blog.csdn.net/leo_888/article/details/88284251 # 启动脚本杀死影响程序的进程 os.system('./killmjpg.sh') time.sleep(0.1) # 要识别的颜色字典 color_dist = { 'red': {'Lower': np.array([156, 60, 60]), 'Upper': np.array([180, 255, 255])}, 'green': {'Lower': np.array([35, 43, 35]), 'Upper': np.array([90, 255, 255])}, 'blue': {'Lower': np.array([110, 80, 80]), 'Upper': np.array([128, 255, 255])}, # 'gray': {'Lower': np.array([0,0,30]),'Upper':np.array([255,40,80]) } }; flag = -1 flag_bak = -1 img_w = 320 img_h = 240 cap = cv2.VideoCapture(0) #打开摄像头 最大范围 640×480 cap.set(3,img_w) #设置画面宽度 cap.set(4,img_h) #设置画面高度 lastTime = time.time() time.sleep(1) while 1: #进入无线循环 global pan,tilt #将摄像头拍摄到的画面作为frame的值 ret,frame = cap.read() for i in color_dist: #1-高斯滤波GaussianBlur() 让图片模糊 frame = cv2.GaussianBlur(frame,(5,5),0) #2-转换HSV的样式 以便检测 hsv = cv2.cvtColor(frame,cv2.COLOR_BGR2HSV) #3-查找字典 mask = cv2.inRange(hsv, color_dist[i]['Lower'], color_dist[i]['Upper']) #4-腐蚀图像 mask = cv2.erode(mask,None,iterations=2) #高斯模糊 mask = cv2.GaussianBlur(mask,(3,3),0) #图像合并 res = cv2.bitwise_and(frame,frame,mask=mask) #6-边缘检测 cnts = cv2.findContours(mask.copy(),cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)[-2] if len(cnts) >0 : #通过边缘检测来确定所识别物体的位置信息得到相对坐标 cnt = max(cnts,key=cv2.contourArea) rect = cv2.minAreaRect(cnt) # 获取最小外接矩形的4个顶点 box = cv2.boxPoints(rect) #获取坐标 长宽 角度 c_x, c_y = rect[0] c_h, c_w = rect[1] c_angle = rect[2] #print(c_x,c_y,c_h,c_w) #判断是否是方块大小且在方框中 if 25 < c_h < 85 and 25 < c_w < 85: # and 50 < c_x < 270 and 60 < c_y < 180: #print('---',c_x,c_y,c_h,c_w) #绘制轮廓 cv2.drawContours(frame, [np.int0(box)], -1, (0, 255, 255), 2) #print(time.time(), 'x=', int(c_x), 'y=', int(c_y), 'c_h=', int(c_h), 'c_w=', int(c_w), 'angle=', int(c_angle)) errorPan = c_x - img_w/2 errorTilt = c_y - img_h/2 #print('errorpan',errorPan) if abs(errorPan)>15: pan=pan-errorPan/75 if abs(errorTilt)>15: tilt=tilt-errorTilt/50 #print('pan',pan) if pan > 180: pan = 180 print("Pan out of Range") if pan < 0: pan = 0 print("pan Out of Range") if tilt > 180: tilt = 180 print("Pan out of Range") if tilt < 0: tilt = 0 print("pan Out of Range") #servo_move(0,180-pan,1,tilt) servo_move(0,pan,1,180-tilt) else: pass cv2.imshow('frame',frame) #将具体的测试效果显示出来 #cv2.imshow('mask',mask) #cv2.imshow('res',res) if cv2.waitKey(5) & 0xFF == 27: #如果按了ESC就退出 当然也可以自己设置 break cap.release() cv2.destroyAllWindows() #后面两句是常规操作,每次使用摄像头都需要这样设置一波
import cv2 #导入库 import numpy as np import time import os from servo_control import * #创建显示控件 from jetcam.utils import bgr8_to_jpeg import traitlets import ipywidgets.widgets as widgets from IPython.display import display disp = widgets.Image(format='jpeg', width=320, height=240) display(disp) #关于色域范围可以百度 HSV #百科：https://baike.baidu.com/item/HSV/547122?fr=aladdin #参考：https://blog.csdn.net/leo_888/article/details/88284251 # 启动脚本杀死影响程序的进程 os.system('./killmjpg.sh') time.sleep(0.1) # 要识别的颜色字典 color_dist = { 'red': {'Lower': np.array([156, 60, 60]), 'Upper': np.array([180, 255, 255])}, 'green': {'Lower': np.array([35, 43, 35]), 'Upper': np.array([90, 255, 255])}, 'blue': {'Lower': np.array([110, 80, 80]), 'Upper': np.array([128, 255, 255])}, # 'gray': {'Lower': np.array([0,0,30]),'Upper':np.array([255,40,80]) } }; flag = -1 flag_bak = -1 img_w = 320 img_h = 240 cap = cv2.VideoCapture(0) #打开摄像头 最大范围 640×480 cap.set(3,img_w) #设置画面宽度 cap.set(4,img_h) #设置画面高度 lastTime = time.time() time.sleep(1) try: while 1: #进入无线循环 global pan,tilt #将摄像头拍摄到的画面作为frame的值 ret,frame = cap.read() for i in color_dist: #1-高斯滤波GaussianBlur() 让图片模糊 frame = cv2.GaussianBlur(frame,(5,5),0) #2-转换HSV的样式 以便检测 hsv = cv2.cvtColor(frame,cv2.COLOR_BGR2HSV) #3-查找字典 mask = cv2.inRange(hsv, color_dist[i]['Lower'], color_dist[i]['Upper']) #4-腐蚀图像 mask = cv2.erode(mask,None,iterations=2) #高斯模糊 mask = cv2.GaussianBlur(mask,(3,3),0) #图像合并 res = cv2.bitwise_and(frame,frame,mask=mask) #6-边缘检测 cnts = cv2.findContours(mask.copy(),cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)[-2] if len(cnts) >0 : #通过边缘检测来确定所识别物体的位置信息得到相对坐标 cnt = max(cnts,key=cv2.contourArea) rect = cv2.minAreaRect(cnt) # 获取最小外接矩形的4个顶点 box = cv2.boxPoints(rect) #获取坐标 长宽 角度 c_x, c_y = rect[0] c_h, c_w = rect[1] c_angle = rect[2] #print(c_x,c_y,c_h,c_w) #判断是否是方块大小且在方框中 if 25 < c_h < 85 and 25 < c_w < 85: # and 50 < c_x < 270 and 60 < c_y < 180: #print('---',c_x,c_y,c_h,c_w) #绘制轮廓 cv2.drawContours(frame, [np.int0(box)], -1, (0, 255, 255), 2) #print(time.time(), 'x=', int(c_x), 'y=', int(c_y), 'c_h=', int(c_h), 'c_w=', int(c_w), 'angle=', int(c_angle)) errorPan = c_x - img_w/2 errorTilt = c_y - img_h/2 #print('errorpan',errorPan) if abs(errorPan)>15: pan=pan-errorPan/75 if abs(errorTilt)>15: tilt=tilt-errorTilt/50 #print('pan',pan) if pan > 180: pan = 180 print("Pan out of Range") if pan < 0: pan = 0 print("pan Out of Range") if tilt > 180: tilt = 180 print("Pan out of Range") if tilt < 0: tilt = 0 print("pan Out of Range") #servo_move(0,180-pan,1,tilt) servo_move(0,pan,1,180-tilt) else: pass disp.value = bgr8_to_jpeg(frame) #cv2.imshow('frame',frame) #将具体的测试效果显示出来 #cv2.imshow('mask',mask) #cv2.imshow('res',res) #if cv2.waitKey(5) & 0xFF == 27: #如果按了ESC就退出 当然也可以自己设置 #break except KeyboardInterrupt: cap.release() #cv2.destroyAllWindows() #后面两句是常规操作,每次使用摄像头都需要这样设置一波
16.人脸跟随
# OpenCV 摄像头云台人脸追踪 import os import cv2 import numpy as np #import Adafruit_PCA9685 import time from servo_control import * # 启动脚本杀死影响程序的进程 os.system('./killmjpg.sh') time.sleep(0.1) servo_move(0,90,1,90) cam = cv2.VideoCapture(0) dis_w = 320 dis_h = 240 cam.set(3, dis_w) cam.set(4, dis_h) # 载入人脸和眼睛的HAAR 模型 face_cascade = cv2.CascadeClassifier('./trains/haarcascade_frontalface_default.xml') eye_cascade = cv2.CascadeClassifier('./trains/haarcascade_eye.xml') def Video_display(): global pan, tilt while True: ret,frame = cam.read() gray = cv2.cvtColor(frame,cv2.COLOR_BGR2GRAY) faces = face_cascade.detectMultiScale(gray,1.3,5) for(x,y,w,h) in faces: cv2.rectangle(frame,(x,y),(x+w,y+h),(0,0,255),2) Xcent = x + w/2 Ycent = y + h/2 errorPan = Xcent - dis_w/2 errorTilt = Ycent - dis_h/2 if abs(errorPan)>15: pan=pan-errorPan/50 if abs(errorTilt)>15: tilt=tilt-errorTilt/50 if pan > 180: pan = 180 print("Pan out of Range") if pan < 0: pan = 0 print("pan Out of Range") if tilt > 180: tilt = 180 print("Pan out of Range") if tilt < 0: tilt = 0 print("pan Out of Range") servo_move(0,pan,1,180-tilt) #servo_move(0,pan,1,180-tilt) roi_gray = gray[y:y+h, x:x+w] roi_color = frame[y:y+h, x:x+w] eyes = eye_cascade.detectMultiScale(roi_gray) for (ex,ey,ew,eh) in eyes: cv2.rectangle(roi_color,(ex,ey),(ex+ew,ey+eh),(0,255,0),2) cv2.imshow('frame', frame) if cv2.waitKey(5) & 0xFF == 27: break if __name__ == '__main__': try: Video_display() except KeyboardInterrupt: cap.release() cv2.destroyAllWindows()
# OpenCV 摄像头云台人脸追踪 import os import cv2 import numpy as np #import Adafruit_PCA9685 import time from servo_control import * #创建显示控件 from jetcam.utils import bgr8_to_jpeg import traitlets import ipywidgets.widgets as widgets from IPython.display import display disp = widgets.Image(format='jpeg', width=320, height=240) display(disp) # 启动脚本杀死影响程序的进程 os.system('./killmjpg.sh') time.sleep(0.1) servo_move(0,90,1,90) cam = cv2.VideoCapture(0) dis_w = 320 dis_h = 240 cam.set(3, dis_w) cam.set(4, dis_h) # 载入人脸和眼睛的HAAR 模型 face_cascade = cv2.CascadeClassifier('./trains/haarcascade_frontalface_default.xml') eye_cascade = cv2.CascadeClassifier('./trains/haarcascade_eye.xml') def Video_display(): global pan, tilt while True: ret,frame = cam.read() gray = cv2.cvtColor(frame,cv2.COLOR_BGR2GRAY) faces = face_cascade.detectMultiScale(gray,1.3,5) for(x,y,w,h) in faces: cv2.rectangle(frame,(x,y),(x+w,y+h),(0,0,255),2) Xcent = x + w/2 Ycent = y + h/2 errorPan = Xcent - dis_w/2 errorTilt = Ycent - dis_h/2 if abs(errorPan)>15: pan=pan-errorPan/50 if abs(errorTilt)>15: tilt=tilt-errorTilt/50 if pan > 180: pan = 180 print("Pan out of Range") if pan < 0: pan = 0 print("pan Out of Range") if tilt > 180: tilt = 180 print("Pan out of Range") if tilt < 0: tilt = 0 print("pan Out of Range") servo_move(0,pan,1,180-tilt) #servo_move(0,pan,1,180-tilt) roi_gray = gray[y:y+h, x:x+w] roi_color = frame[y:y+h, x:x+w] eyes = eye_cascade.detectMultiScale(roi_gray) for (ex,ey,ew,eh) in eyes: cv2.rectangle(roi_color,(ex,ey),(ex+ew,ey+eh),(0,255,0),2) disp.value = bgr8_to_jpeg(frame) #cv2.imshow('frame', frame) #if cv2.waitKey(5) & 0xFF == 27: #break if __name__ == '__main__': try: Video_display() except KeyboardInterrupt: cam.release()
import time import Adafruit_PCA9685 pwm = Adafruit_PCA9685.PCA9685() pwm.set_pwm_freq(50) def servo_move(basechannel,basevalue,tiltchannel,tiltvalue): basevalue=4096*((basevalue*11)+500)/20000 pwm.set_pwm(basechannel,0,int(basevalue)) tiltvalue=4096*((tiltvalue*11)+500)/20000 pwm.set_pwm(tiltchannel,0,int(tiltvalue)) pan = 90 tilt = 90 servo_move(0, pan, 1, tilt) ''' if __name__ == '__main__': servo_move(0,30,1,30) time.sleep(3) for angle in range(180, 0, -5): servo_move(0, angle, 1, angle) time.sleep(0.5) servo_move(0, pan, 1, tilt) ''' #print('end')
浙公网安备 33010602011771号