OpenCV obtains the mobile camera and performs screen recording

Posted by DaiWelsh on Fri, 25 Feb 2022 14:59:21 +0100

1. Download Mobile Software

You need to download the IP camera software on the mobile phone. You can turn the mobile device into a wireless IP camera with two-way audio support through the built-in RTSP and HTTP server and use it for security monitoring. You can use the browser on the computer to view it. Here I use OpenCV to read the video stream and record the screen

2. Get video stream address

Open the next good software and click "open IP camera server" to obtain the address of the video stream under the current intranet. Here, I choose the first video stream of RTSP protocol. When accessing, I need to enter the account password, which is admin by default


3. Code writing

3.1 reading video stream

Create a new VideoCapture object and read the video stream

cap = cv2.VideoCapture("rtsp://admin:admin@******.local:8554/live")

3.2 display video

Use the while statement to cycle through the video and display it

while True:
    success,img =

3.3 recording video

To create a VideoWriter object and write an image into a video, you need to specify several parameters

  1. Represents the new file stored in the read video frame
  2. It refers to the coding format of video storage
  3. Represents the number of frames per second
  4. Represents the length, width and size of the image
fourcc = cv2.VideoWriter_fourcc(*'mp4v')
size = (int(self.capture.get(cv2.CAP_PROP_FRAME_WIDTH)), \
out = cv2.VideoWriter("name.mp4", fourcc, 25, self.size)
out.write(img)    #Write pictures to video

4. Effect display

5. Code display

According to the above ideas, I have encapsulated and optimized the code. By using the space bar to start and end the recording, press ESC to exit the program, and the current recording state and current time are displayed in the video. Normal indicates the normal display state, Recoding indicates that it is recording, and the space bar can be used to switch the state. If you use ESC to launch the program without pressing the pause key, you can also save the video normally.

import cv2 
import time
class Camera:
    def __init__(self, addr, save_dir="./"):
        """initial value"""
        self.save_dir = save_dir
        self.addr = addr
        self.isRecoding = False
        self.capture =cv2.VideoCapture(self.addr)
        self.fourcc = cv2.VideoWriter_fourcc(*'mp4v')
        self.size = (int(self.capture.get(cv2.CAP_PROP_FRAME_WIDTH)), \
    def exit(self):
        if self.isRecoding:

    def mark(self,img):
        now = int(time.time())
        timeArray = time.localtime(now)
        if self.isRecoding:
            state = "Recoding:"
            state = "Normal:"
        otherStyleTime = state + time.strftime("%Y-%m-%d %H:%M:%S", timeArray)
        otherStyleTime = otherStyleTime.encode("gbk").decode(errors="ignore")
        cv2.putText(img, otherStyleTime, (50, 50), cv2.FONT_HERSHEY_SIMPLEX, 1, (55,255,155), 2)
        return img

    def open(self):
        """open camera"""
        while (self.capture.isOpened()):
            success,img =
            if success:
                key = cv2.waitKey(5)
                if key == 32:       #Press the spacebar to start and end recording
                    self.isRecoding = not self.isRecoding
                    if self.isRecoding:
                        vedeo_name = time.strftime("%Y%m%d%H%M%S", time.localtime(int(time.time()))) + ".mp4"
                        self.out = cv2.VideoWriter(self.save_dir+vedeo_name, self.fourcc, 25, self.size)
                img = self.mark(img)
                result = cv2.cvtColor(img, cv2.COLOR_RGB2BGR)
                if self.isRecoding:
                if key == 27:       #ESC key exit
                print("Fail to get image")
if __name__ == '__main__':
    addr = "rtsp://admin:admin@zzttekiiPhone.local:8554/live"
    cam1 = Camera(addr)

6. Download address

The file has been uploaded to Gitee and can be downloaded if necessary

Topics: OpenCV Computer Vision