Develop Reference

tkinter + OpenCV WebCam Very Slow Video Stream

I am using tkinter and opencv for the first time and have successfully built a GUI for my project, however, I cannot figure out why my video stream is updating so extremely slow. I am grabbing frames very quickly but it seems that the update on the screen gets exponentially slower. I am seeing somewhere around 30 seconds of lag when I first launch the program but it eventually slows to a halt. I am connecting to three cameras but only displaying one at a time. The cameras all display and the selection buttons work. My only issue is the display refresh rate.
This is running in Python3.7 on a raspberry pi4. I can connect to the camera via web browser and it appears to have no lag.
I have been searching for answers but cannot seem to find anything that helps. Can anyone offer some help with this?
Here's my program (I have removed unrelated code):
#!/usr/bin/env python3
import time
from tkinter import *
import cv2
from PIL import Image, ImageTk
#GUI
class robotGUI:
def __init__(self):
self.selectedCam = "front"
self.window = Tk()
#Setup the window to fit the Raspberry Pi Touch Display = 800x400 and align top left
self.window.geometry("800x480+0+0")
self.window.overrideredirect(True)
self.window.fullScreenState = False
#Create Frame for Video Window
self.videoFrame = Frame(self.window, relief=SUNKEN, bd=2)
self.videoFrame.place(x=0, y=0, height=457, width=650)
#Create the Video Window
self.video = Label(self.videoFrame, bd=0, relief=FLAT, width=644, height=451)
self.video.place(x=0, y=0)
self.vid = VideoCapture()
self.camUpdateFreq = 250
self.updateCams()
#Create the Button Frame
self.buttonFrame = Frame(self.window, relief=FLAT)
self.buttonFrame.place(x=651, y=0, height=457, width=149)
#Create Buttons
#Select Front Camera Button
self.frontCamButton = Button(self.buttonFrame, text="Front Camera", command=lambda: self.selectCam("front"))
self.frontCamButton.place(x=24, y=50, height=30, width=100)
#Select Boom Camera Button
self.boomCamButton = Button(self.buttonFrame, text="Boom Camera", command=lambda: self.selectCam("boom"))
self.boomCamButton.place(x=24, y=130, height=30, width=100)
#Select Rear Camera Button
self.rearCamButton = Button(self.buttonFrame, text="Rear Camera", command=lambda: self.selectCam("rear"))
self.rearCamButton.place(x=24, y=210, height=30, width=100)
#Close Button
self.exitButton = Button(self.buttonFrame, text="Close", command=self.window.destroy)
self.exitButton.place(x=24, y=400, height=30, width=100)
#Start the main loop for the gui
self.window.mainloop()
def selectCam(self, cam):
if (cam.lower() == "front"):
self.selectedCam = "front"
self.statusBarLeft['text'] = "Front Camera Selected"
elif (cam.lower() == "boom"):
self.selectedCam = "boom"
self.statusBarLeft['text'] = "Boom Camera Selected"
elif (cam.lower() == "rear"):
self.selectedCam = "rear"
self.statusBarLeft['text'] = "Rear Camera Selected"
def updateCams(self):
#Get a frame from the selected camera
ret, frame = self.vid.get_frame(self.selectedCam)
if ret:
imageCV2 = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
img = Image.fromarray(imageCV2)
imgPhoto = ImageTk.PhotoImage(image=img)
self.video.imgPhoto = imgPhoto
self.video.configure(image=imgPhoto)
self.window.after(self.camUpdateFreq, self.updateCams)
#Video Camera Class
class VideoCapture:
def __init__(self):
#Define Locals
FrontCameraAddress = "rtsp://admin:password#192.168.5.20:8554/12"
BoomCameraAddress = "rtsp://admin:password#192.168.5.21:8554/12"
RearCameraAddress = "rtsp://admin:password#192.168.5.22:8554/12"
#Open Front Video Camera Source
self.vidFront = cv2.VideoCapture(FrontCameraAddress)
self.vidBoom = cv2.VideoCapture(BoomCameraAddress)
self.vidRear = cv2.VideoCapture(RearCameraAddress)
#Verify that the Camera Streams Opened
if not self.vidFront.isOpened():
raise ValueError("Unable to open video source to Front Camera")
if not self.vidBoom.isOpened():
raise ValueError("Unable to open video source to Boom Camera")
if not self.vidRear.isOpened():
raise ValueError("Unable to open video source to Rear Camera")
#Get One Frame from the Selected Camera
def get_frame(self, camera="front"):
#Attempt to Get Front Camera Frame
if (camera.lower() == "front"):
#If Stream Still Open Return a Frame
if self.vidFront.isOpened():
ret, frame = self.vidFront.read()
if ret:
#Return a boolean success flag and the current frame converted to BGR
return (ret, frame)
else:
return (ret, None)
else:
return (ret, None)
#Attempt to Get Boom Camera Frame
elif (camera.lower() == "boom"):
#If Stream Still Open Return a Frame
if self.vidBoom.isOpened():
ret, frame = self.vidBoom.read()
if ret:
#Return a boolean success flag and the current frame converted to BGR
return (ret, frame)
else:
return (ret, None)
else:
return (ret, None)
#Attempt to Get Rear Camera Frame
elif (camera.lower() == "rear"):
#If Stream Still Open Return a Frame
if self.vidRear.isOpened():
ret, frame = self.vidRear.read()
if ret:
#Return a boolean success flag and the current frame converted to BGR
return (ret, frame)
else:
return (ret, None)
else:
return (ret, None)
else:
return (False, None)
#Release the video sources when the object is destroyed
def __del__(self):
if self.vidFront.isOpened():
self.vidFront.release()
if self.vidBoom.isOpened():
self.vidBoom.release()
if self.vidRear.isOpened():
self.vidRear.release()
#Main Routine - Only run if called from main program instance
if __name__ == '__main__':
try:
#Create GUI Object
app = robotGUI()
except Exception as e:
print("Exception: " + str(e))
finally:
print("Cleaning Up")
NOTE: In this copy of the program, I am updating every 250ms but I have tried smaller numbers down to around 3 but the frames still seem to be behind. Is there a better way to do this?
NOTE2: After working with this more today, I realize that openCV is definitely buffering frames for each camera starting when the cv2.VideoCapture() function is called for each camera. The read() function does seem to be pulling the next frame from the buffer which explains why it is taking so long to update and why the image I see on the screen never catches up to reality. I changed my test code to only connect to one camera at a time and use the cv2.release() function any time I am not actively viewing a camera. This improved things quite a bit. I also set the update function to run every 1ms and I am using the grab() function to grab a frame every cycle but I am only processing and displaying every 10th cycle which has also improved some. I still have some lag that I would love to remove if anyone has any suggestions.
My RTSP stream shows with zero noticeable lag when viewed in a web browser. Does anyone know how I can get the same effect in tkinter? I am not married to openCV.

How to fix problem with save a video frame by frame

The problem is that I try to save a video file I get a one frame each time from another function (I check this is not the same frame...), the video created but only with one frame. I run with a loop outside the class Video_utility and send frame to the function save_and_display_video.
import cv2
class Video_utility:
def __init__(self, name_video, format_video, display, fps, size):
self.name_video = name_video
self.format_video = format_video
self.display = display
self.fps = fps
self.size = size
self.stream_frame = None
self.flag_update = True
self.display = True
self.fourcc = cv2.VideoWriter_fourcc(*'XVID')
self.out = cv2.VideoWriter(name_video, self.fourcc, fps, self.size)
self.i = 0
def save_and_display_video(self, frame):
frame = cv2.resize(frame,(self.size))
self.out.write(frame)
self.out.release()
cv2.destroyAllWindows()

Don't close the file after every frame
You have self.out.release() at your save_and_display_video() function.
You'd need to do that only after you've received the whole video.

PIL: Jump to next photo after one click event

I am preparing data for deep running. So I have to get certain pixel coordinates for each picture. Only one coordinate per photo is required.
So when I use PIL to input one click, I try to implement coordinates so that I can go to the next picture.
However, when I write the code as below, the coordinates are output in duplicate to only one picture, and the next picture does not appear on the screen.
How can I make sure that only one coordinate is recorded on a single picture?
from PIL import Image, ImageTk
import tkinter
import os
URL = './SavedImage/'
imgList = os.listdir(URL)
print(imgList)
width = 852
height = 480
stepW = 852/4
stepH = 480/5
def callback(event):
print("clicked at: ", event.x, event.y)
window = tkinter.Tk(className='pla')
for file in sorted(imgList):
a=True
image = Image.open(os.path.join(URL, file))
print(image)
canvas = tkinter.Canvas(window, width=image.size[0], height=image.size[1])
canvas.pack()
image_tk = ImageTk.PhotoImage(image)
canvas.create_image(image.size[0]//2, image.size[1]//2, image=image_tk)
canvas.bind("<Button-1>", callback)
tkinter.mainloop()

I am not 100% sure I understand what you need but it looks like to me you are trying to get one set of cord's for each image in a list of images.
I would do this by creating a function and a tracking variable to loop through each image on at a time and on click update a new list with the image and cord's then loop to next image.
Let me know if you have any questions.
Example:
from PIL import Image, ImageTk
import tkinter
import os
URL = './SavedImage/'
imgList = os.listdir(URL)
width = 852
height = 480
stepW = 852/4
stepH = 480/5
tracker = 0
list_images_with_cords = [] # added list for final results
def callback(event):
# Added global's.
global tracker,list_images_with_cords
# Used to append final results to list.
list_images_with_cords.append([imgList[tracker], event.x, event.y])
# This tracker lets us go through each item on the list.
tracker += 1
# After appending list go to next image.
open_next()
window = tkinter.Tk(className='pla')
# Creates just one canvas that we can update later.
canvas = tkinter.Canvas(window)
canvas.pack()
def open_next():
# Adding global's.
global image, canvas, image_tk, tracker
# Clearing canvas before drawing new image.
canvas.delete("all")
# Checking for valid index in list.
if tracker < len(imgList):
image = Image.open(os.path.join(URL, imgList[tracker]))
# use config() to update canvas.
canvas.config(width=image.size[0], height=image.size[1])
image_tk = ImageTk.PhotoImage(image)
canvas.create_image(image.size[0]//2, image.size[1]//2, image=image_tk)
canvas.bind("<Button-1>", callback)
else:
# This else statement is just for when we run out of images.
# It will display all the results in a textbox.
canvas.destroy()
txt = tkinter.Text(window, width=25)
txt.pack()
for item in list_images_with_cords:
txt.insert("end", "{}\n\n".format(item))
open_next()
tkinter.mainloop()

Play an Animated GIF in python with tkinter

I am wanting to create a virtual pet style game using python3 and tkinter. So far I have the main window and have started putting labels in, but the issue I am having is playing an animated gif. I have searched on here and have found some answers, but they keep throwing errors. The result I found has the index position of the gif using PhotoImage continue through a certain range.
# Loop through the index of the animated gif
frame2 = [PhotoImage(file='images/ball-1.gif', format = 'gif -index %i' %i) for i in range(100)]
def update(ind):
frame = frame2[ind]
ind += 1
img.configure(image=frame)
ms.after(100, update, ind)
img = Label(ms)
img.place(x=250, y=250, anchor="center")
ms.after(0, update, 0)
ms.mainloop()
When I run this in terminal with "pyhton3 main.py" I get the following error:
_tkinter.TclError: no image data for this index
What am I overlooking or completely leaving out?
Here is the link to the GitHub repository to see the full project:VirtPet_Python
Thanks in advance!

The error means that you tried to load 100 frames, but the gif has less than that.
Animated gifs in tkinter are notoriously bad. I wrote this code an age ago that you can steal from, but will get laggy with anything but small gifs:
import tkinter as tk
from PIL import Image, ImageTk
from itertools import count
class ImageLabel(tk.Label):
"""a label that displays images, and plays them if they are gifs"""
def load(self, im):
if isinstance(im, str):
im = Image.open(im)
self.loc = 0
self.frames = []
try:
for i in count(1):
self.frames.append(ImageTk.PhotoImage(im.copy()))
im.seek(i)
except EOFError:
pass
try:
self.delay = im.info['duration']
except:
self.delay = 100
if len(self.frames) == 1:
self.config(image=self.frames[0])
else:
self.next_frame()
def unload(self):
self.config(image="")
self.frames = None
def next_frame(self):
if self.frames:
self.loc += 1
self.loc %= len(self.frames)
self.config(image=self.frames[self.loc])
self.after(self.delay, self.next_frame)
root = tk.Tk()
lbl = ImageLabel(root)
lbl.pack()
lbl.load('ball-1.gif')
root.mainloop()

First of all, you need to know what is the last range of your GIF file. so by changing the different value of i, you will get it.For my condition is 31.
then just need to put the condition.So it will play gif infinitely.
from tkinter import *
import time
import os
root = Tk()
frames = [PhotoImage(file='./images/play.gif',format = 'gif -index %i' %(i)) for i in range(31)]
def update(ind):
frame = frames[ind]
ind += 1
print(ind)
if ind>30: #With this condition it will play gif infinitely
ind = 0
label.configure(image=frame)
root.after(100, update, ind)
label = Label(root)
label.pack()
root.after(0, update, 0)
root.mainloop()

A very simple approach would be to use multithreading.
To run the GIF infinitely in a Tkinter window you should follow the following:
Create a function to run the GIF.
Put your code to run the GIF inside while True inside the function.
Create a thread to run the function.
Run root.mainloop() in the primary flow of the program.
Use time.sleep() to control the speed of your animation.
Refer to my code below:
i=0
ph = ImageTk.PhotoImage(Image.fromarray(imageframes[i]))
imglabel=Label(f2,image=ph)
imglabel.grid(row=0,column=0)
def runthegif(root,i):
while True:
i = i + 7
i= i % 150
ph=ImageTk.PhotoImage(PhotoImage(file='images/ball.gif',format='gif -index %i' %i))
imagelabel=Label(f2,image=ph)
imagelabel.grid(row=0,column=0)
time.sleep(0.1)
t1=threading.Thread(target=runthegif,args=(root,i))
t1.start()
root.mainloop()

How to display a animated GIF in a Tkinter Label - Python [duplicate]

This question already has answers here:
Play an Animated GIF in python with tkinter
(3 answers)
Closed 5 years ago.
I've been trying to play an animated gif using Tkinter.PhotoImage, but haven't been seeing any success. It displays the image, but not the animation. The following is my code:
root = Tkinter.Tk()
photo = Tkinter.PhotoImage(file = "path/to/image.gif")
label = Tkinter.Label(image = photo)
label.pack()
root.mainloop()
It displays the image in a window, and that's it. I'm thinking that the issue has something to do with Tkinter.Label but I'm not sure. I've looked for solutions but they all tell me to use PIL (Python Imaging Library), and it's something that I don't want to use.
With the answer, I created some more code (which still doesn't work...), here it is:
from Tkinter import *
def run_animation():
while True:
try:
global photo
global frame
global label
photo = PhotoImage(
file = photo_path,
format = "gif - {}".format(frame)
)
label.configure(image = nextframe)
frame = frame + 1
except Exception:
frame = 1
break
root = Tk()
photo_path = "/users/zinedine/downloads/091.gif"
photo = PhotoImage(
file = photo_path,
)
label = Label(
image = photo
)
animate = Button(
root,
text = "animate",
command = run_animation
)
label.pack()
animate.pack()
root.mainloop()
Thanks for everything! :)

You have to drive the animation yourself in Tk. An animated gif consists of a number of frames in a single file. Tk loads the first frame but you can specify different frames by passing an index parameter when creating the image. For example:
frame2 = PhotoImage(file=imagefilename, format="gif -index 2")
If you load up all the frames into separate PhotoImages and then use timer events to switch the frame being shown (label.configure(image=nextframe)). The delay on the timer lets you control the animation speed. There is nothing provided to give you the number of frames in the image other than it failing to create a frame once you exceed the frame count.
See the photo Tk manual page for the official word.

Here's a simpler example without creating an object:
from tkinter import *
import time
import os
root = Tk()
frameCnt = 12
frames = [PhotoImage(file='mygif.gif',format = 'gif -index %i' %(i)) for i in range(frameCnt)]
def update(ind):
frame = frames[ind]
ind += 1
if ind == frameCnt:
ind = 0
label.configure(image=frame)
root.after(100, update, ind)
label = Label(root)
label.pack()
root.after(0, update, 0)
root.mainloop()