I need a function that changes their size for a given number of photos, use all available computing powers. I found that I can use function parallel for this, but I don't know how to use it properly.
Here is my code:
from random import randint
from fastai.core import parallel
import cv2
def image_conversion(list, size, directory, conversion):
if conversion == '--resize':
for img in list:
imageAsNumpy = cv2.imread(img)
dim = (size[1], size[0])
resized = cv2.resize(imageAsNumpy, dim, interpolation = cv2.INTER_AREA)
writeStatus = cv2.imwrite(directory + img, resized)
if writeStatus is True:
print('Imagine written successfully!')
else:
print('Something went wrong!')
elif conversion == '--random crop':
for img in list:
image = cv2.imread(img)
height, width = image.shape[:2]
h = randint(0, height)
w = randint(0, width)
cropped = image[h:h+width, w:w+height]
status = cv2.imwrite(directory + img, cropped)
if status is True:
print('Imagine cropped')
else:
print('Problem')
But when I tried to call parallel function it fails every time, for example the progress bar goes to a 100%, but with none of the work is completed. Could someone please help me understand what I am doing wrong?
Related
I've created a simple Python application that uses the CV2 computer vision library to recognise a template image on a webpage.
I give the application a template image that it needs to recognise on the source image. In this case, the source image is a screenshot of the website www.google.com and the template image is the Google search button.
Template image
I thought the application worked at first, but it's drawing the rectangle completely in the wrong place on the input (source) image. I've added a picture below of where the application located the template image.
Result
Here's the source code.
Main Application Source
import cv2
import numpy
from io import BytesIO
from PIL import Image
import matplotlib.pyplot as plt
import numpy as np
class Automate:
def __init__(self):
chrome_options = Options()
chrome_options.add_argument("kiosk")
self.driver = webdriver.Chrome(ChromeDriverManager("93.0.4577.63").install(), options=chrome_options)
#self.driver = webdriver.Chrome(executable_path='./chromedriver',options=chrome_options)
self.screenShot = None
self.finalImage = None
def open_webpage(self, url):
print(f"Open webpage {url}")
self.driver.get(url)
def close_webpage(self):
Event().wait(5)
self.driver.close()
print("Closing webpage")
def snap_screen(self):
print("Capturing screen")
self.screenShot = "screenshot.png"
self.driver.save_screenshot(self.screenShot)
print("done.")
def match(self, image, template):
# convert images to greyscale.
src = cv2.cvtColor(cv2.imread(image), cv2.COLOR_BGR2GRAY)
temp = cv2.cvtColor(cv2.imread(template), cv2.COLOR_BGR2GRAY)
cv2.imshow("out", temp)
cv2.waitKey(0)
height, width = src.shape
H, W = temp.shape
result = cv2.matchTemplate(src, temp, cv2.cv2.TM_CCOEFF_NORMED)
minVal, maxVal, minLoc, maxLoc = cv2.minMaxLoc(result)
location = maxLoc
bottomRight = (location[0] + W, location[1] + H)
src2 = cv2.imread(image)
cv2.rectangle(src2, location, bottomRight, (0, 0, 255), 5)
cv2.imshow("output", src2)
cv2.waitKey(0)
cv2.destroyAllWindows()
def main():
url = "http://www.google.com"
auto = Automate()
auto.open_webpage(url)
auto.snap_screen()
auto.close_webpage()
match_image = "images/templates/google-button.png"
# Match screenshot with template image.
auto.check_match(
image=auto.screenShot,
template=match_image
)
I'd appreciate any help or advice on how to solve this issue.
Update
Following the advice given by user zteffi, I resized my template image to the correct image dimensions. After doing this, the match template function works as expected.
You want to make sure that your template image is a close as possible to the actual size of the image you want to be located in the base image. In my case, this was around 150 x 150 or 200 x 200 so that it will be easier to find the button.
I'm trying to save image buffers as .pngs with blender's python API (but this is mainly just a python question)
I'm trying to speed things up by making a thread to save the image, and the function that creates the thread is being called from a callback-handler that is activated whenever the 3D screen is refreshed, here is the full code (its a little messy):
import base64, io, os, bgl, gpu, bpy, threading, time, sys
import numpy as np
from gpu_extras.presets import draw_texture_2d
from PIL import Image
import multiprocessing.pool as mpool
finalPath = bpy.context.scene.render.filepath + "hithere.png"
WIDTH = 1920
HEIGHT = 1080
offscreen = gpu.types.GPUOffScreen(WIDTH, HEIGHT)
def draw2():
global finalPath
global array
global WIDTH
global HEIGHT
global needsSaving
context = bpy.context
scene = context.scene
view_matrix = scene.camera.matrix_world.inverted()
projection_matrix = scene.camera.calc_matrix_camera(
context.depsgraph, x=WIDTH, y=HEIGHT)
offscreen.draw_view3d(
scene,
context.view_layer,
context.space_data,
context.region,
view_matrix,
projection_matrix)
bgl.glDisable(bgl.GL_DEPTH_TEST)
draw_texture_2d(offscreen.color_texture, (0, -125), WIDTH, HEIGHT)
buffer = bgl.Buffer(bgl.GL_BYTE, WIDTH * HEIGHT * 4)
bgl.glReadBuffer(bgl.GL_BACK)
bgl.glReadPixels(0, -125, WIDTH, HEIGHT, bgl.GL_RGBA, bgl.GL_UNSIGNED_BYTE, buffer)
needle = threading.Thread(target=saveIt,args=[buffer, finalPath, WIDTH, HEIGHT])
needle.daemon = True
needle.start()
#### thread.start_new_thread(saveIt,(buffer, finalPath, WIDTH, HEIGHT))
def coby(scene):
frame = scene.frame_current
folder = scene.render.filepath
myFormat = "png"#scene.render.image_settings.renderformat.lower()
outputPath = os.path.join(folder, "%05d.%s" % (frame, myFormat))
global finalPath
finalPath = outputPath
h = bpy.types.SpaceView3D.draw_handler_add(draw2, (), 'WINDOW', 'POST_PIXEL')
bpy.app.handlers.frame_change_pre.clear()
bpy.app.handlers.frame_change_pre.append(coby)
def saveIt(buffer, path, width, height):
array = np.asarray(buffer, dtype=np.uint8)
myBytes = array.tobytes()
im = Image.frombytes("RGBA",(width, height), myBytes)
rawBytes = io.BytesIO()
im.save(rawBytes, "PNG")
rawBytes.seek(0)
base64Encoded = base64.b64encode(rawBytes.read())
txt = "data:image/png;base64," + base64Encoded.decode()
f = open(finalPath, "wb")
f.write(base64.decodebytes(base64Encoded))
f.close()
it actually does work, except when I play the timeline in blender (which calls the frame_pre callback and also the 3D view-refresh callback (not sure in which order though), most of my images are replaced, except some are not, as can be seen in this screenshot:
[![enter image description here][1]][1]
I originally had all blue-planed images, then I ran the script through the thread, and it replaced almost all of them, except some of the blue-planed images still remain (seemingly at a random interval). This worked fine if I call .join() right after I make the thread, or don't use the thread at all, but seemingly the thread is the only way to make it work a lot faster.
I've been looking around for how to use threads with queues and pooling (How to use python multiprocessing Pool.map within loop, What happened to thread.start_new_thread in python 3,
How can I make a background, non-blocking input loop in python?, Creating Threads in python),
SO: Why aren't all of the threads finishing?
[1]: https://i.stack.imgur.com/nJCwH.png .................
I'm trying to save a captured 640x480 RGB image with NAO's front camera to my computer. I'm using python and PIL to do so. Unfortunately, the image just won't save on my computer, no matter what image type or path I use for the parameters of the Image.save()- Method. the image created with PIL contains valid RGB-information though. Here's my code sample from choregraphe:
import Image
def onInput_onStart(self):
cam_input = ALProxy("ALVideoDevice")
nameId = cam_input.subscribeCamera("Test_Cam", 1, 2, 13, 20)
image = cam_input.getImageRemote(nameId) #captures an image
w = image[0] #get the image width
h = image[1] #get the image height
pixel_array = image[6] #contains the image data
result = Image.fromstring("RGB", (w, h), pixel_array)
#the following line doesnt work
result.save("C:\Users\Claudia\Desktop\NAO\Bilder\test.png", "PNG")
cam_input.releaseImage(nameId)
cam_input.unsubscribe(nameId)
pass
Thank you so much for your help in advance!
- a frustrated student
In the comment, you say the code is pasted from choregraphe, so I guess you launch it using choregraphe.
If so, then the code is injected into your robot then started.
So your image is saved to the NAO hard drive and I guess your robot doesn't have a folder named: "C:\Users\Claudia\Desktop\NAO\Bilder\test.png".
So change the path to "/home/nao/test.png", start your code, then log into your NAO using putty or browse folder using winscp (as it looks like you're using windows).
And you should see your image-file.
In order for your code to run correctly it needs to be properly indented. Your code should look like this:
import Image
def onInput_onStart(self):
cam_input = ALProxy("ALVideoDevice")
nameId = cam_input.subscribeCamera("Test_Cam", 1, 2, 13, 20)
image = cam_input.getImageRemote(nameId) #captures an image
w = image[0] #get the image width
h = image[1] #get the image height
pixel_array = image[6] #contains the image data
...
Make sure to indent everything that's inside the def onInput_onStart(self): method.
Sorry for the late response, but it maybe helpful for someone. You should try it with naoqi. Here is the documentation for retriving images
http://doc.aldebaran.com/2-4/dev/python/examples/vision/get_image.html
The original code was not working for me so I made some tweeks.
parser = argparse.ArgumentParser()
parser.add_argument("--ip", type=str, default="nao.local.",
help="Robot IP address. On robot or Local Naoqi: use
'nao.local.'.")
parser.add_argument("--port", type=int, default=9559,
help="Naoqi port number")
args = parser.parse_args()
session = qi.Session()
try:
session.connect("tcp://" + args.ip + ":" + str(args.port))
except RuntimeError:
pass
"""
First get an image, then show it on the screen with PIL.
"""
# Get the service ALVideoDevice.
video_service = session.service("ALVideoDevice")
resolution = 2 # VGA
colorSpace = 11 # RGB
videoClient = video_service.subscribe("python_client",0,3,13,1)
t0 = time.time()
# Get a camera image.
# image[6] contains the image data passed as an array of ASCII chars.
naoImage = video_service.getImageRemote(videoClient)
t1 = time.time()
# Time the image transfer.
print ("acquisition delay ", t1 - t0)
#video_service.unsubscribe(videoClient)
# Now we work with the image returned and save it as a PNG using ImageDraw
# package.
# Get the image size and pixel array.
imageWidth = naoImage[0]
imageHeight = naoImage[1]
array = naoImage[6]
image_string = str(bytearray(array))
# Create a PIL Image from our pixel array.
im = Image.fromstring("RGB", (imageWidth, imageHeight), image_string)
# Save the image.
im.save("C:\\Users\\Lenovo\\Desktop\\PROJEKTI\\python2-
connect4\\camImage.png", "PNG")
Be careful to use Python 2.7.
The code runs on your computer not the NAO robot!
I am attempting to write a function which imports a specified CDF datafile, formats the image as a numpy array and does some slight refinements on the image to remove background.This works fine, however I encounter an error when I try to define the name of the array as one of my parameters in my function:
from netCDF4 import Dataset
import numpy as np
def importfunction(datafile, imagelabel):
f = Dataset(datafile)
locationfloatfield = f.variables['FloatField']
floatfield = locationfloatfield[:]
img = floatfield.flatten()
img = scipy.signal.detrend(img)
imagelabel = np.reshape(img, (256, 256))
imagelabel += abs(imagelabel.min())
imagelabel *= (1.0/imagelabel.max())
I attempt to label the array as imagelabel which is specified when I call the function. However when I call this function i get this error:
importfunction('..../CS191mk2153-M-Xp-Topo.nc', label)
NameError: name 'label' is not defined
I am unsure how to fix this
Maybe what you want to do is:
def importfunction(datafile):
f = Dataset(datafile)
locationfloatfield = f.variables['FloatField']
floatfield = locationfloatfield[:]
img = floatfield.flatten()
img = scipy.signal.detrend(img)
imagelabel = np.reshape(img, (256, 256))
imagelabel += abs(imagelabel.min())
imagelabel *= (1.0/imagelabel.max())
return imagelabel
then call:
label = importfunction('..../CS191mk2153-M-Xp-Topo.nc')
Imagine a red circle with a black dropshadow that fades away on top of a fully transparent background. When I open and resave the image with PIL the background remains fully transparent but the dropshadow becomes full black.
The problem appears without even altering the image:
image = Image.open('input.png')
image = image.convert('RGBA')
image.save('output.png')
I want to keep the image looking exactly as the original so that I can crop or resize it.
EDIT: Here's a PNG that demonstrates the effect. It was converted to 8bit by using PNGNQ.
When using the above Python code it comes out as the following:
It looks like PIL currently doesn't support full alpha for PNG8.
There is a patch here for read-only support: http://mail.python.org/pipermail/image-sig/2010-October/006533.html
If you're feeling naughty, you could monkeypatch PIL:
from PIL import Image, ImageFile, PngImagePlugin
def patched_chunk_tRNS(self, pos, len):
i16 = PngImagePlugin.i16
s = ImageFile._safe_read(self.fp, len)
if self.im_mode == "P":
self.im_info["transparency"] = map(ord, s)
elif self.im_mode == "L":
self.im_info["transparency"] = i16(s)
elif self.im_mode == "RGB":
self.im_info["transparency"] = i16(s), i16(s[2:]), i16(s[4:])
return s
PngImagePlugin.PngStream.chunk_tRNS = patched_chunk_tRNS
def patched_load(self):
if self.im and self.palette and self.palette.dirty:
apply(self.im.putpalette, self.palette.getdata())
self.palette.dirty = 0
self.palette.rawmode = None
try:
trans = self.info["transparency"]
except KeyError:
self.palette.mode = "RGB"
else:
try:
for i, a in enumerate(trans):
self.im.putpalettealpha(i, a)
except TypeError:
self.im.putpalettealpha(trans, 0)
self.palette.mode = "RGBA"
if self.im:
return self.im.pixel_access(self.readonly)
Image.Image.load = patched_load
Image.open('kHrY6.png').convert('RGBA').save('kHrY6-out.png')
I think that the problem has been somewhat resolved, but is it possible that you need to set the depth of the alpha channel?