We Keep Coding

Issues multi-trackers and opencv's version

I'm trying to implement the OpenCV's multi-tracker as explained in this tutorial (https://www.learnopencv.com/multitracker-multiple-object-tracking-using-opencv-c-python/) but I'm facing some issues with the version of OpenCV.
As told in other answers on SO, for the single and multi tracker constructors I have to use opencv-contrib; however, once installed opencv-contrib in place of opencv, I'm able no more to use the VideoCapture function. I'm in a deadlock since I need both functions but contrib-opencv and opencv seem cannot co-exist in the same interpreter.
Has anyone faced the same issue and can help me with this?
import sys
import cv2
from random import randint
trackerTypes = ['BOOSTING', 'MIL', 'KCF', 'TLD', 'MEDIANFLOW', 'GOTURN', 'MOSSE', 'CSRT']
def createTrackerByName(trackerType):
# Create a tracker based on tracker name
if trackerType == trackerTypes[0]:
tracker = cv2.TrackerBoosting_create()
elif trackerType == trackerTypes[1]:
tracker = cv2.TrackerMIL_create()
elif trackerType == trackerTypes[2]:
tracker = cv2.TrackerKCF_create()
elif trackerType == trackerTypes[3]:
tracker = cv2.TrackerTLD_create()
elif trackerType == trackerTypes[4]:
tracker = cv2.TrackerMedianFlow_create()
elif trackerType == trackerTypes[5]:
tracker = cv2.TrackerGOTURN_create()
elif trackerType == trackerTypes[6]:
tracker = cv2.TrackerMOSSE_create()
elif trackerType == trackerTypes[7]:
tracker = cv2.TrackerCSRT_create()
else:
tracker = None
print('Incorrect tracker name')
print('Available trackers are:')
for t in trackerTypes:
print(t)
return tracker
def main():
# Set video to load
videoPath = "../../data/videos/case-1-2D.avi"
# Create a video capture object to read videos
cap = cv2.VideoCapture(videoPath)
# Read first frame
success, frame = cap.read()
# quit if unable to read the video file
if not success:
print('Failed to read video')
sys.exit(1)
## Select boxes
bboxes = []
colors = []
while True:
bbox = cv2.selectROI('MultiTracker', frame)
bboxes.append(bbox)
colors.append((randint(0, 255), randint(0, 255), randint(0, 255)))
print("Press q to quit selecting boxes and start tracking")
print("Press any other key to select next object")
k = cv2.waitKey(0) & 0xFF
if (k == 113): # q is pressed
break
print('Selected bounding boxes {}'.format(bboxes))
# Specify the tracker type
trackerType = "CSRT"
# Create MultiTracker object
multiTracker = cv2.MultiTracker_create()
# Initialize MultiTracker
for bbox in bboxes:
multiTracker.add(createTrackerByName(trackerType), frame, bbox)
# Process video and track objects
while cap.isOpened():
success, frame = cap.read()
if not success:
break
# get updated location of objects in subsequent frames
success, boxes = multiTracker.update(frame)
# draw tracked objects
for i, newbox in enumerate(boxes):
p1 = (int(newbox[0]), int(newbox[1]))
p2 = (int(newbox[0] + newbox[2]), int(newbox[1] + newbox[3]))
cv2.rectangle(frame, p1, p2, colors[i], 2, 1)
# show frame
cv2.imshow('MultiTracker', frame)
# quit on ESC button
if cv2.waitKey(1) & 0xFF == 27: # Esc pressed
break
if __name__ == '__main__':
main()
If I use opencv-python package I get the error:
AttributeError: module 'cv2.cv2' has no attribute
'MultiTracker_create'
If I use opencv-contrib-python package I get the error:
AttributeError: module 'cv2' has no attribute
'VideoCapture'
If it can help, I'm working with PyCharm on MacOS using a VirtualEnv for my Python interpreter.
I already tried to installed opencv-contrib (alone), opencv (alone), keep both opencv-contrib and opencv on my interpreter, downgrade opencv to an early version that should support the functions I need (opencv 3.4).

Python select ROI OpenCV

Sample Image
Hello,
I created an application in python that select the Region of Interest(ROI) of an image, record and label it. But I has a limit of one ROI per image, anyone know how to have multiple selection of ROI per image? Also on attached image, as you can see I have multiple window, I want it to be in one window with different options, what packages are use on this kind of application.
here's my code in python using opencv2. Thank you in advance for the help
for image in filelist:
img = cv2.imread(image)
fromCenter = False
r = cv2.selectROI(img, fromCenter)
lbl = simpledialog.askstring("Image Label", "Please Enter Label")
result = eTree.SubElement(results, "Image")
path = eTree.SubElement(result, 'Path')
roi = eTree.SubElement(result, 'ROI')
label = eTree.SubElement(result, 'Label')
path.text = str(image)
roi.text = str(r)
label.text = str(lbl)
tree = eTree.ElementTree(results)
i = i + 1
if i == count:
format = [('XML Files', '*.xml'), ('All Files', '*.*')]
save = filedialog.asksaveasfilename(filetype=format, defaultextension='*.xml')
tree.write(save, xml_declaration=True, encoding='utf-8', method="xml")

Well at least for the first part of the question, have you considered to try the cv2.createROIs() instead of cv2.createROI() ? When the image window is opened you then select your first ROI and press enter, then the second and press enter etc. And when you are finished then press the escape key. It returns x,y,w,h of each ROI. Note that you will have to change your code accordingly but it will allow you to select multiple ROI.
Input image:
Example:
import cv2
img = cv2.imread('rois.png')
fromCenter = False
ROIs = cv2.selectROIs('Select ROIs', img, fromCenter)
ROI_1 = img[ROIs[0][1]:ROIs[0][1]+ROIs[0][3], ROIs[0][0]:ROIs[0][0]+ROIs[0][2]]
ROI_2 = img[ROIs[1][1]:ROIs[1][1]+ROIs[1][3], ROIs[1][0]:ROIs[1][0]+ROIs[1][2]]
ROI_3 = img[ROIs[2][1]:ROIs[2][1]+ROIs[2][3], ROIs[2][0]:ROIs[2][0]+ROIs[2][2]]
cv2.imshow('1', ROI_1)
cv2.imshow('2', ROI_2)
cv2.imshow('3', ROI_3)
cv2.waitKey(0)
cv2.destroyAllWindows()
Result:

For custom ROI you can use EasyROI. It supports rectangle, line, circle and polygon.
For using it:
pip install EasyROI
from EasyROI import EasyROI
roi_helper = EasyROI()
roi = roi_helper.draw_rectangle(frame, quantity=2)

openCL cv2.UMat object is not iterable

well actually i dont really understand how to use this T-API opencl and still newbie at it, in the documentation
https://www.learnopencv.com/opencv-transparent-api/
it use cv2.UMat to the .image file and then read it. In my problem i want to use the opencl T-API to my reconizer.predict line because the image is taken/processing while streaming the camera
recognizer = cv2.face.LBPHFaceRecognizer_create()
#colec = cv2.face.MinDistancePredictCollector()
recognizer.read("trainer_data_array.yml")
labels = {"persons_name":0}
with open("labels.pickle", "rb") as f:
og_labels = pickle.load(f)
labels = {v:k for k,v in og_labels.items()}
cap = cv2.VideoCapture(0)
while(True):
#video cap
ret, frame = cap.read()
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
faces = face_cascade.detectMultiScale(gray, scaleFactor=1.5, minNeighbors=5)
for (x,y,w,h) in faces:
#print(x,y,w,h)
roi_gray = gray[y:y+h, x:x+w]
roy_color = frame[y:y+h, x:x+w]
#recognize how?
id_ , conf = recognizer.predict(roi_gray) #some error, some say cuz its opencv 3.1.0 bug
#solution : up opencv to 3.3 or just use MinDistancePredictCollector(...)
if conf>=45 and conf<=85:
print(idppl)
print(labels[idppl])
font = cv2.FONT_HERSHEY_SIMPLEX
name = labels[idppl]
color = (255,255,255)
stroke = 2
cv2.putText(frame,name,(x,y),font,1,color,stroke,cv2.LINE_AA)
elif conf > 85:
print("unknown")
can someone help me how to do it, if just put it raw like id_ , conf = cv2.UMat(recognizer.predict(roi_gray))
it give me error cv2.UMat' object is not iterable
well without the T-API mybe this line program still give a good frame rate, but after given many modification or implementation/process it'll run in low frame rate when detecting/recog- people face.
this why i want to use openCl so when it run in gpu mybe will give me a pretty good frame rate

Accessing Beaglebone python opencv usb camera but the display showing black screen

I am facing issue while accessing USB camera using beagle-bone black wireless.
Firstly the error is "select timeout" exception which was resolved by this post
Now I am facing the black screen in output.
Here is the testing code I am using.
from cv2 import *
# initialize the camera
cam = VideoCapture(0) # 0 -> index of camera
print "Cam capture"
cam.set(3,320)
cam.set(4,240)
print "Cam set"
s, img = cam.read()
print "Cam read"
if s: # frame captured without any errors
namedWindow("cam-test",CV_WINDOW_AUTOSIZE)
imshow("cam-test",img)
while True:
key = waitKey(30)
if key == ord('q') :
destroyWindow("cam-test")
I have already check that video0 in /dev directory.

The issue is that you need to call 'cam.read()andimshow()` inside the while loop
What you're doing is that you're reading just the first frame, then showing it, and you whileloop isn't doing anything. When the camera boots, the first frame is just a blank screen , which is what you see.
The code should be more like:
while True:
s, img = cam.read()
imshow("cam-test",img)
key = waitKey(30)
if key == ord('q') :
destroyWindow("cam-test")

Python PIL - All areas of PNG with opacity > 0 have their opacity set to 1

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?