我正在研究Yolov3，OpenCV，python和Flask。在这里，我将详细解释每个项目的用法。
Yolov3-检测和识别视频输入中的对象
Opencv-捕获视频输入中检测到的对象的图像。
Flask-作为Web服务器，因为它支持python语言。
我的目标 开发能够捕获对象图像并直接或实时在flask Web中更新的应用程序
供您参考，目前，我的系统能够使用OpenCv和python捕获图像并将其保存为一个名为图像的文件夹。您可以参考下面的代码。
opencv和python代码以捕获检测到的对象的图像
for i in range(len(boxes)):
if i in indexes:
x,y,w,h = boxes[i]
label = str(LABELS[class_ids[i]])
confidence= confidences[i]
color = colors[class_ids[i]]
crop_img = image[y:y + h, x:x + w]
imagesPath = "images/file_%d.jpg"%self.d
cv2.imwrite(imagesPath, crop_img)
self.d+=1
cv2.rectangle(image,(x,y),(x+w,y+h),color,2)
cv2.putText(image,label+" "+str(round(confidence,2)),(x,y+30),font,1,(255,255,255),2)
当图像成功保存在图像文件夹中时。我将所有图像都转换为base64并在flask中作为json返回。在这里，我将python代码附加在flask中。
将图像转换为base64并在html模板中呈现
@app.route("/images")
def get_images():
directory = os.listdir('C:/Users/HP/Miniconda3/envs/count_vechicle/coding/images')
os.chdir('C:/Users/HP/Miniconda3/envs/count_vechicle/coding/images')
image_list= list()
for file in directory:
data = dict()
base = os.path.basename(file)
data["label"] = base
open_file = open(file,'rb')
image_read = open_file.read()
image_64_encode = base64.encodebytes(image_read)
data["data"] = image_64_encode.decode('ascii')
image_list.append(data)
final_data = {'files':image_list}
return render_template('images.html', final_data=final_data)
images.html
<!DOCTYPE html>
<html>
<head>
<!DOCTYPE html>
<html lang="en">
<head>
<title>yolo</title>
</head>
<body>
<h1 class="logo">Results</h1>
<ul>
{% for data in final_data.files %}
<li>{{data.label}}</li>
<img alt="embedded" src="data:image/jpg;base64,{{data.data}}"/>
{% endfor %}
</ul>
</body>
</html>
问题和疑问
我这里有两个主要问题。
1）当我运行我的应用程序时，我的应用程序很好地捕获了图像。但是，当打开image.html页面以显示图片时，它会自动停止。为什么会发生以及如何解决呢？
2）如果应用程序即使我也可以捕获图像，则打开image.html页面。我应该怎么做才能直接或实时更新我的网站？
在这里附上完整的代码。因为我已经工作了很多周，但仍然找不到解决方案。希望有人可以帮忙。让我知道您是否需要更多信息
代码
App.py
from flask import Flask, render_template, Response,jsonify
from camera import VideoCamera
import numpy as np
import os
import time
import detect as dt
from PIL import Image
import cv2
import base64
import json
from pprint import pprint
app = Flask(__name__)
def success_handle(output, status=200, mimetype='application/json'):
return Response(output, status=status, mimetype=mimetype)
@app.route("/images")
def get_images():
directory = os.listdir('C:/Users/HP/Miniconda3/envs/count_vechicle/coding/images')
os.chdir('C:/Users/HP/Miniconda3/envs/count_vechicle/coding/images')
flist = list()
for file in directory:
data = dict()
base = os.path.basename(file)
data["label"] = base
open_file = open(file,'rb')
image_read = open_file.read()
image_64_encode = base64.encodebytes(image_read)
data["data"] = image_64_encode.decode('ascii')
flist.append(data)
final_data = {'files':flist}
return render_template('images.html', final_data=final_data)
@app.route('/')
def index():
return render_template('index.html')
def gen(camera):
frame_id = 0
while True:
frame = camera.get_frame()
yield (b'--frame\r\n'
b'Content-Type: image/jpeg\r\n\r\n' + frame + b'\r\n\r\n')
@app.route('/video_feed')
def video_feed():
return Response(gen(VideoCamera()),
mimetype='multipart/x-mixed-replace; boundary=frame')
if __name__ == '__main__':
app.run(host='0.0.0.0', debug=True)
camera.py
import cv2
import time
import os
import numpy as np
font = cv2.FONT_HERSHEY_PLAIN
starting_time= time.time()
frame_id = 0
count = 0
d = 0
labelsPath = os.path.sep.join(["yolo-coco", "coco.names"])
weightsPath = os.path.sep.join(["yolo-coco", "yolov3.weights"])
configPath = os.path.sep.join(["yolo-coco", "yolov3.cfg"])
#imagesPath = 'C:/Users/HP/Miniconda3/envs/count_vechicle/coding/images/file_%d.jpg"%d'
labelsPath = os.path.sep.join(["yolo-coco", "coco.names"])
VideoPath = os.path.sep.join(["videos", "highway.mp4"])
LABELS = open(labelsPath).read().strip().split("\n")
net = cv2.dnn.readNet(configPath, weightsPath)
layer_names = net.getLayerNames()
outputlayers = [layer_names[i[0] - 1] for i in net.getUnconnectedOutLayers()]
colors= np.random.uniform(0,255,size=(len(LABELS),3))
class VideoCamera(object):
def __init__(self):
# Using OpenCV to capture from device 0. If you have trouble capturing
# from a webcam, comment the line below out and use a video file
# instead.
self.video = cv2.VideoCapture(VideoPath)
self.frame_id = 0
self.d = 0
# If you decide to use video.mp4, you must have this file in the folder
# as the main.py.
# self.video = cv2.VideoCapture('video.mp4')
def __del__(self):
self.video.release()
def get_frame(self):
success, image = self.video.read()
# We are using Motion JPEG, but OpenCV defaults to capture raw images,
# so we must encode it into JPEG in order to correctly display the
# video stream.
self.frame_id+=1
#print(frame_id)
height,width,channels = image.shape
#print (frame.shape)
#detecting objects
blob = cv2.dnn.blobFromImage(image,0.00392,(320,320),(0,0,0),True,crop=False) #reduce 416 to 320
net.setInput(blob)
outs = net.forward(outputlayers)
#print(outs)
print(outs[1])
#Showing info on screen/ get confidence score of algorithm in detecting an object in blob
class_ids=[]
confidences=[]
boxes=[]
for out in outs:
#print(out)
for detection in out:
scores = detection[5:]
class_id = np.argmax(scores)
confidence = scores[class_id]
print(confidence)
if confidence > 0.8:
#object detected
center_x= int(detection[0]*width)
center_y= int(detection[1]*height)
w = int(detection[2]*width)
h = int(detection[3]*height)
#cv2.circle(img,(center_x,center_y),10,(0,255,0),2)
#rectangle co-ordinaters
x=int(center_x - w/2)
y=int(center_y - h/2)
#cv2.rectangle(img,(x,y),(x+w,y+h),(0,255,0),2)
boxes.append([x,y,w,h]) #put all rectangle areas
confidences.append(float(confidence)) #how confidence was that object detected and show that percentage
class_ids.append(class_id) #name of the object tha was detected
indexes = cv2.dnn.NMSBoxes(boxes,confidences,0.4,0.6)
# result = open('C:/Users/HP/Miniconda3/envs/count_vechicle/coding/images/frame%04d.txt'%(count), 'w')
for i in range(len(boxes)):
if i in indexes:
x,y,w,h = boxes[i]
label = str(LABELS[class_ids[i]])
# cv2.imwrite(label, crop_img)
confidence= confidences[i]
color = colors[class_ids[i]]
crop_img = image[y:y + h, x:x + w]
imagesPath = "images/file_%d.jpg"%self.d
cv2.imwrite(imagesPath, crop_img)
self.d+=1
cv2.rectangle(image,(x,y),(x+w,y+h),color,2)
cv2.putText(image,label+" "+str(round(confidence,2)),(x,y+30),font,1,(255,255,255),2)
elapsed_time = time.time() - starting_time
fps=frame_id/elapsed_time
cv2.putText(image,"FPS:"+str(round(fps,2)),(10,50),font,2,(0,0,0),1)
# cv2.imshow("Image",image)
# key = cv2.waitKey(1) #wait 1ms the loop will start again and we will process the next frame
# if key == 27: #esc key stops the process
# break;
ret, jpeg = cv2.imencode('.jpg', image)
return jpeg.tobytes()
index.html
<html>
<head>
<title>Object Detection</title>
</head>
<body>
<h1>Video Streaming Demonstration</h1>
<img src="{{ url_for('video_feed') }}">
</body>
</html>
images.html
<!DOCTYPE html>
<html>
<head>
<!DOCTYPE html>
<html lang="en">
<head>
<title>yolo</title>
</head>
<body>
<h1 class="logo">Results</h1>
<ul>
{% for data in final_data.files %}
<li>{{data.label}}</li>
<img alt="embedded" src="data:image/jpg;base64,{{data.data}}"/>
{% endfor %}
</ul>
</body>
</html>