i'm trying to take pictures with OpenCV 4.4.0.40 and only save them if a switch, read by an Arduino, is press.
So far everything work, but it's super slow, it take about 15 seconds for the Switch value to change.
Arduino = SerialObject()
if os.path.exists(PathCouleur) and os.path.exists(PathGris):
Images = cv2.VideoCapture(Camera)
Images.set(3, 1920)
Images.set(4, 1080)
Images.set(10, Brightness)
Compte = 0
SwitchNumero = 0
while True:
Sucess, Video = Images.read()
cv2.imshow("Camera", Video)
Video = cv2.resize(Video, (Largeur, Hauteur))
Switch = Arduino.getData()
try:
if Switch[0] == "1":
blur = cv2.Laplacian(Video, cv2.CV_64F).var()
if blur < MinBlur:
cv2.imwrite(PathCouleur + ".png", Video)
cv2.imwrite(PathGris + ".png", cv2.cvtColor(Video, cv2.COLOR_BGR2GRAY))
Compte += 1
except IndexError as err:
pass
if cv2.waitKey(40) == 27:
break
Images.release()
cv2.destroyAllWindows()
else:
print("Erreur, aucun Path")
the saved images width are 640 and the height is 480 and the showimage is 1920x1080 but even without the showimage it's slow.
Can someone help me optimize this code please?
I would say that this snippet of code is responsible for the delay, since you have two major calls that depend on external devices to respond (the OpenCV calls and Arduino.getData()).
Sucess, Video = Images.read()
cv2.imshow("Camera", Video)
Video = cv2.resize(Video, (Largeur, Hauteur))
Switch = Arduino.getData()
One solution that comes to mind is to use the multithreading lib and separate the Arduino.getData() call from the main loop cycle and use it in a separate class to run in the background (you should use a sleep on the secondary thread, something like 50 or 100ms delay).
That way you should have a better response on the Switch event, when it tries to read the value on the main loop.
cam=cv2.VideoCapture(0,cv2.CAP_DSHOW)
use cv2.CAP_DSHOW it will improve the loading time of camera becasue it wil give you video feed directly
I've just begun using the OpenCV library for Python and came across something I didn't understand.
cap = cv2.VideoCapture(0)
while True:
ret, frame = cap.read() #returns ret and the frame
cv2.imshow('frame',frame)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
When I use cv2.waitKey(1), I get a continuous live video feed from my laptops webcam. However when I use cv2.waitKey(0), I get still images. Every time I close the window, another one pops up with another picture taken at the time.
Why does it not show as a continuous feed?
From the doc:
1.waitKey(0) will display the window infinitely until any keypress (it is suitable for image display).
2.waitKey(1) will display a frame for 1 ms, after which display will be automatically closed. Since the OS has a minimum time between switching threads, the function will not wait exactly 1 ms, it will wait at least 1 ms, depending on what else is running on your computer at that time.
So, if you use waitKey(0) you see a still image until you actually press something while for waitKey(1) the function will show a frame for at least 1 ms only.
waitKey(0) will pause your screen because it will wait infinitely for keyPress on your keyboard and will not refresh the frame(cap.read()) using your WebCam.
waitKey(1) will wait for keyPress for just 1 millisecond and it will continue to refresh and read frame from your webcam using cap.read().
More clearly, Use debugger in your code.When using waitKey(0) in the while loop, the debugger never crosses this statement and does not refresh the frame and hence the frame output seems stable.Does not move.
Where as with waitKey(1), the debugger will cross the code after pausing at
if cv2.waitKey(1) & 0xFF == ord('q')
for 1 milli second.
From the documentation you can see that cv2.waitKey(delay) waits for delay milliseconds if delay is positive but forever (waits for a key event infinitely) if it's zero or negative. That's why you see these differences in behavior.
In the case of cv2.waitKey(1) this is, in fact, negligible but its use provides the user the opportunity to press a key (the key might be caught in some next iteration but does not make a big difference).
Delay in milliseconds. 0 is the special value that means “forever”.
The function waitKey waits for a key event infinitely (when \texttt{delay}\leq 0 ) or for delay milliseconds, when it is positive. Since the OS has a minimum time between switching threads, the function will not wait exactly delay ms, it will wait at least delay ms, depending on what else is running on your computer at that time. It returns the code of the pressed key or -1 if no key was pressed before the specified time had elapsed.
Note: This function is the only method in HighGUI that can fetch and handle events, so it needs to be called periodically for normal event processing unless HighGUI is used within an environment that takes care of event processing.
Note: The function only works if there is at least one HighGUI window created and the window is active. If there are several HighGUI windows, any of them can be active
cv2.waitkey(1) is being used in while loop.It shows the output for 1msec but because of infinite while loop it is the sequence of images that are perceived by our brain as a single continuos video.
Hope this helped.
Indratej Reddy's answer is the only one to clearly explain why this works — and why no GUI events act until waitkey is called. Unclear how to handle menu/mouse events — e.g., ignores menu quit.
$ python3
>>> import cv2
>>> im = cv2.imread("/tmp/portrait.jpg")
>>> print("im =", im)
im = […]
>>> cv2.imshow("image", im)
>>> print("imshow called")
imshow called
>>> print("waitKey =", cv2.waitKey(0))
waitKey = 32
>>> print("destroyAllWindows =", cv2.destroyAllWindows())
destroyAllWindows = None
>>> print("Final waitKey 1ms =", cv2.waitKey(1))
Final waitKey 1ms = -1
>>>
I'm working with Python OpenCV to a project that as an initial step involves capturing an image from a webcam; I tried to automate this process by using capture = cv2.VideoCapture and capture.read(), but the camera's video mode activation and its subsequent self-adjusting are too slow for what I want to achieve in the end.
Is there a more direct method of automatically capturing a screenshot with Python (and OpenCV)? If not, do you have any alternative suggestion?
Thanks
If you want your camera screenshot function to be responsive, you need to initialize the camera capture outside of this function.
On the following code snippet, the screenshot function is triggered by pressing c:
import cv2
def screenshot():
global cam
cv2.imshow("screenshot", cam.read()[1]) # shows the screenshot directly
#cv2.imwrite('screenshot.png',cam.read()[1]) # or saves it to disk
if __name__ == '__main__':
cam = cv2.VideoCapture(0) # initializes video capture
while True:
ret, img = cam.read()
cv2.imshow("cameraFeed", img) # a window is needed as a context for key capturing (here, I display the camera feed, but there could be anything in the window)
ch = cv2.waitKey(5)
if ch == 27:
break
if ch == ord('c'): # calls screenshot function when 'c' is pressed
screenshot()
cv2.destroyAllWindows()
To clarify: cameraFeed window is only here for the purpose of the demo (where screenshot is triggered manually). If screenshot is called automatically in your program, then you don't need this part.
Hope it helps!
Basically you need to do 3 things:
#init the cam
video_capture = cv2.VideoCapture(0)
#get a frame from cam
ret, frame = video_capture.read()
#write that to disk
cv2.imwrite('screenshot.png',frame)
of course, you should wait a while before, if not you could save a weird black screen (or just the 1st thing the camera got :-) )
The code
import cv
capture = cv.CaptureFromFile("a.avi")
while True:
frame = cv.QueryFrame(capture)
cv.ShowImage("a',frame)
Shows the same initial frame from the video repeatedly (QueryFrame is not advancing the video and grabbing the next frame). It works fine if the video is captured from a webcam.
Any ideas?
I see the same mistakes over and over again, so this is probably the last time I'll address them. Hopefully people will start using the search box in the future and dig a little deeper.
Call cv.WaitKey() after displaying the frame. If don't have a delay between displaying the frames some problems could happen. I believe this the problem.
Code defensively: if you are calling a function/method that can fail, believe in Murphy, and add the appropriate check to verify it doesn't:
import cv
capture = cv.CaptureFromFile("a.avi")
if not capture :
print "Error loading video file"
# Should exit the application
while True:
frame = cv.QueryFrame(capture)
if not frame:
print "Could not retrieve frame"
cv.ShowImage("a", frame)
k = cv.WaitKey(10)
if k == 27:
break # ESC key was pressed
My program generates a series of windows using the following code:
def display(img, name, fun):
global clicked
cv.NamedWindow(name, 1)
cv.ShowImage(name, img)
cv.SetMouseCallback(name, fun, img)
while cv.WaitKey(33) == -1:
if clicked == 1:
clicked = 0
cv.ShowImage(name, img)
cv.DestroyWindow(name)
I press "q" within the gui window to close it. However, the code continues to the next call of the display function and displays a second gui window while not closing the first. I'm using a Mac with OpenCV 2.1, running the program in Terminal. How can I close the gui windows? Thanks.
You need to run cv.startWindowThread() after opening the window.
I had the same issue and now this works for me.
Hope this helps for future readers. And there is also a cv2 binding (I advise to use that instead of cv).
This code works for me:
import cv2 as cv
import time
WINDOW_NAME = "win"
image = cv.imread("ela.jpg", cv.CV_LOAD_IMAGE_COLOR)
cv.namedWindow(WINDOW_NAME, cv.CV_WINDOW_AUTOSIZE)
initialtime = time.time()
cv.startWindowThread()
while (time.time() - initialtime < 5):
print "in first while"
cv.imshow(WINDOW_NAME, image)
cv.waitKey(1000)
cv.waitKey(1)
cv.destroyAllWindows()
cv.waitKey(1)
initialtime = time.time()
while (time.time() - initialtime < 6):
print "in second while"
The same issue happens with the C++ version, on Linux:
Trying to close OpenCV window has no effect
There are a few peculiarities with the GUI in OpenCV. The destroyImage call fails to close a window (atleast under Linux, where the default backend was Gtk+ until 2.1.0) unless waitKey was called to pump the events. Adding a waitKey(1) call right after destroyWindow may work.
Even so, closing is not guaranteed; the the waitKey function is only intercepted if a window has focus, and so if the window didn't have focus at the time you invoked destroyWindow, chances are it'll stay visible till the next destroyWindow call.
I'm assuming this is a behaviour that stems from Gtk+; the function didn't give me any trouble when I used it under Windows.
Sayem2603
I tried your solution and it worked for me - thanks! I did some trial and error and discovered that looping 4 times did the trick for me... or posting the same code 4 times just the same..
Further, I drilled down to:
cv2.destroyAllWindows()
cv2.waitKey(1)
cv2.waitKey(1)
cv2.waitKey(1)
cv2.waitKey(1)
or simply calling DestroyAllWindows and then looping the waitKey() code 4 times:
cv2.destroyAllWindows()
for i in range (1,5):
cv2.waitKey(1)
Worked as well. I am not savvy enough to know why this works exactly, though I assume it has something to do with the interruption and delay created by looping that code(?)
Matthäus Brandl said, above, that the third waitKey() worked for him, so perhaps it is slightly different on each system? (I am running Linux Mint with 3.16.1 kernel and python 2.7)
The delay, alone, doesn't explain it, as simply increasing the delay time on the waitKey() does not do the trick. (Also looped print("Hello") 1000 times instead of using wiatKey() just to see if the delay that created helped any - it did not.) Must have something more to do with how waitKey() interacts with window events.
OpenCV Docs say: "This function is the only method in HighGUI that can fetch and handle events, so it needs to be called periodically for normal event processing unless HighGUI is used within an environment that takes care of event processing."
Perhaps it creates an interrupt of sorts in the GUI display that allows the destroyAllWindows() action to process?
J
Here is what worked for me:
cv2.namedWindow("image")
cv2.imshow('image', img)
cv2.waitKey(0) # close window when a key press is detected
cv2.destroyWindow('image')
cv2.waitKey(1)
This solution works for me (under Ubuntu 12.04 with python open in the shell):
Re-invoke cv.ShowImage after the window is 'destroyed'.
If you are using Spyder ( Anaconda Package ) there is the problem.
None of the solutions worked for me.
I discovered that the problem wasn't the functions, but a problem on Spyder really. Try to use a texteditor plus running on terminal and you be fine using simply:
WINDOW_NAME = "win"
image = cv.imread("foto.jpg", 0)
cv.namedWindow(WINDOW_NAME, cv.CV_WINDOW_AUTOSIZE)
cv.startWindowThread()
cv.imshow(WINDOW_NAME, image)
cv.waitKey()
cv.destroyAllWindows()
I solved the problem by calling cv2.waitKey(1) in a for loop, I don't know why it worked but gets my job done, so I didn't bother myself further.
for i in range(1,10):
cv2.destroyAllWindows()
cv2.waitkey(1)
you are welcome to explain.
It seems that none of the above solutions worked for me if I run it on Jupyter Notebook (the window hangs when closing and you need to force quit Python to close the window).
I am on macOS High Sierra 10.13.4, Python 3.6.5, OpenCV 3.4.1.
The below code works if you run it as a .py file (source: https://www.learnopencv.com/read-write-and-display-a-video-using-opencv-cpp-python/). It opens the camera, records the video, closes the window successfully upon pressing 'q', and saves the video in .avi format.
import cv2
import numpy as np
# Create a VideoCapture object
cap = cv2.VideoCapture(0)
# Check if camera opened successfully
if (cap.isOpened() == False):
print("Unable to read camera feed")
# Default resolutions of the frame are obtained.The default resolutions are system dependent.
# We convert the resolutions from float to integer.
frame_width = int(cap.get(3))
frame_height = int(cap.get(4))
# Define the codec and create VideoWriter object.The output is stored in 'outpy.avi' file.
out = cv2.VideoWriter('outpy.avi',cv2.VideoWriter_fourcc('M','J','P','G'), 10, (frame_width,frame_height))
while(True):
ret, frame = cap.read()
if ret == True:
# Write the frame into the file 'output.avi'
out.write(frame)
# Display the resulting frame
cv2.imshow('frame',frame)
# Press Q on keyboard to stop recording
if cv2.waitKey(1) & 0xFF == ord('q'):
break
# Break the loop
else:
break
# When everything done, release the video capture and video write objects
cap.release()
out.release()
# Closes all the frames
cv2.destroyAllWindows()
Fiddling around with this issue in the python console I observed the following behavior:
issuing a cv2.imshow after cv2.destroyWindow sometimes closes the window. Albeit the old window pops up again with the next highgui call, e.g., cv2.namedWindow
the third call of cv2.waitKey after cv2.destroyWindow closed the window every time I tried. Additionally the closed window remained closed, even when using cv2.namedWindow afterwards
Hope this helps somebody.
(I used Ubuntu 12.10 with python 2.7.3 but OpenCV 2.4.2 from the 13.04 repos)
After searching aroung for some time, none of the solutions provided worked for me so since there's a bug in this function and I did not have time to fix it, I did not have to use the cv2 window to show the frames. Once a few frames have been saved, you can open the file in a different viewer, like VLC or MoviePlayer ( for linux ).
Here's how i did mine.
import cv2
threadDie = True # change this to false elsewhere to stop getting the video
def getVideo(Message):
print Message
print "Opening url"
video = cv2.VideoCapture("rtsp://username:passwordp#IpAddress:554/axis-media/media.amp")
print "Opened url"
fourcc = cv2.cv.CV_FOURCC('X','V','I','D')
fps = 25.0 # or 30.0 for a better quality stream
writer = cv2.VideoWriter('out.avi', fourcc,fps, (640,480),1)
i = 0
print "Reading frames "
while threadDie:
ret, img = video.read()
print "frame number: ",i
i=i+1
writer.write(img)
del(video)
print "Finished capturing video"
Then open the file with a different viewer, prabably in a nother function, like if you like vlc, you can start it and pass the saved file as a parameter. On the terminal, i would do this
vlc out.avi #out.avi is my video file being saved by the function above.
This worked for me on arch linux.
I had the same issue. The problem is that while(cap.isOpened()): loop does not finish so that I added below structure. When video has no frame in the following part, it returns ret values as False. Normally, I put destroyAllWindows command out of loop but I moved it into the loop. It works in my code properly.
while(cap.isOpened()):
ret, frame = cap.read()
if ret == False:
cap.release()
cv2.waitKey(1)
cv2.destroyAllWindows()
cv2.waitKey(1)
This worked for me in spyder :
import cv2 as cv
cv.namedWindow("image")
img = cv.imread("image_name.jpg")
cv.imshow("image",img)
cv.waitKey(5000) # 5 sec delay before image window closes
cv.destroyWindow("image")
Remember use only cv.waitKey(positive Integer) for this to work
cv2.imshow("the image I want to show ",img)
cv2.waitKey(0)
cv2.destroyAllWindows()
cv2.waitKey(1) # to close the window.
The above code worked well for me.
I'm using Mac and python 3.7 .
Close the terminal, later close the window, it worked for me in Visual Studio Code in Windows; I made a task to compile and run the executable in the terminal, the program used my webcam to capture video and display it in a QT window, when I clicked the close button it didn't close, it reopened itself again and continued with the program until I closed the terminal and later could close the program window without it reopening again.