The Code:
I am attempting to construct a python script that takes the RGB values of an image, along with the contours width and height into a neural network to determine if the little rc car I have constructed should go forward or turn right.
Code:
import cv2
cam = cv2.VideoCapture(2)
class Pic_Capture:
def __init__(self):
# print("Fuck you")
self.i = 0
def Img_Mods(self, cam):
self.ret, self.frame = cam.read()
self.buf1 = cv2.flip(self.frame, 1)
# self.buf2 = cv2.cvtColor(self.buf1, cv2.COLOR_BGR2RGB)
self.buf3 = cv2.GaussianBlur(self.buf1,(5,5),0)
cv2.imshow('buf2', self.buf3)
def Measurement_Bullshit(self):
self.buf4 = cv2.cvtColor(self.buf3, cv2.COLOR_RGB2GRAY)
self.buf5 = cv2.adaptiveThreshold(self.buf4,255,cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY,11,2)
_, contours, hierarchy = cv2.findContours(self.buf5, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE)
for cnt in contours:
M = cv2.moments(cnt)
self.cx = int(M['m10']/M['m00'])
self.cy = int(M['m01']/M['m00'])
print(self.cx, self.cy)
cv2.imshow('buf4', self.buf4)
running = True
i = 0
Camera = Pic_Capture()
while running:
Camera.Img_Mods(cam)
if i % 30 == 0 and i != 0:
Camera.Off_and_On_Again()
Camera.Measurement_Bullshit()
i += 1
print(i)
if cv2.waitKey(1) == 27:
Camera.Stop_Code()
Question 1:
I consistently receive this within my console and I would like to get to rid of it for debugging later down the line.
Corrupt JPEG data: 4 extraneous bytes before marker 0xd5
Every time this pops up it is at a different marker, so I do not know how to not print that to the console.
Question 2:
After the code within the if statement: if i % 10 == 0 and i != 0: runs the code spits back this error.
File "Nueral_Network_Creation.py", line 23, in Measurement_Bullshit self.cx = int(M['m10']/M['m00']) ZeroDivisionError: float division by zero
I understand that it is most likely that the code is not finding enough contours to be run properly or maybe it is something completely different. As always any help is much appreciated.
Related
The problem is, program is really lengthy and computationally expensive.
so is there any way to make this program faster or any other way to write this code?
I am beginner in python and would love to take all suggestions or different approach then this program
also i am new to the stack overflow so if anything is wrong in this post or any issue in program please point out in comment .
first section of code is
#TEST V2.1 multitracker
import cv2
import numpy as np
#path = (input("enter the video path: "))
cap = cv2.VideoCapture(" YOUR VIDEO PATH ")
# creating the dictionary to add all the wanted trackers in OpenCV that can be used for tracking in future
OBJECT_TRACKING_MACHINE = {
"csrt": cv2.legacy.TrackerCSRT_create,
"kcf": cv2.legacy.TrackerKCF_create,
"boosting": cv2.legacy.TrackerBoosting_create,
"mil": cv2.legacy.TrackerMIL_create,
"tld": cv2.legacy.TrackerTLD_create,
"medianflow": cv2.legacy.TrackerMedianFlow_create,
"mosse": cv2.legacy.TrackerMOSSE_create
}
# Creating the MultiTracker variable object to store the
trackers = cv2.legacy.MultiTracker_create()
here I started the loop
while True:
frame = cap.read()[1]
#print("freame start",frame)
if frame is None:
print("error getting the video,please check the input")
break
frame = cv2.resize(frame,(1080,720))
Thresh = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
#Thresh = cv2.adaptiveThreshold(gray, 185, cv2.ADAPTIVE_THRESH_MEAN_C,cv2.THRESH_BINARY_INV, 11, 6)
#print(trackers.update(Thresh))
(success, boxes) = trackers.update(Thresh)
# loop over the bounding boxes and draw them on the frame
if success == False:
bound_boxes = trackers.getObjects()
idx = np.where(bound_boxes.sum(axis= 1) != 0)[0]
bound_boxes = bound_boxes[idx]
trackers = cv2.legacy.MultiTracker_create()
for bound_box in bound_boxes:
trackers.add(tracker,Thresh,bound_box)
x,y,w,h = cv2.boundingRect(Thresh)
k = cv2.waitKey(50)
And I am guessing this is the section which is making the program slow
if is there any different way to represent this part or any idea different then this
for i,box in enumerate(boxes):
(x, y, w, h) = [int(v) for v in box]
#cv2.rectangle(Thresh, (x, y), (x + w, y + h), (255, 255, 255), 2)
#cv2.putText(Thresh,('TRACKING BOX NO-'+str(i)),(x+10,y-3),cv2.FONT_HERSHEY_PLAIN,1.0,(255,255,0),2)
arr = boxes.astype(int)
if i == 0 :
Roi = Thresh[(arr[i,1]):(arr[i,1]+arr[i,3]),(arr[i,0]):(arr[i,0]+arr[i,2])]
murg = cv2.resize(Roi,(300,200))
cv2.imshow("horizon", murg)
#print(murg)
if i == 1 :
Roi1 = Thresh[(arr[i,1]):(arr[i,1]+arr[i,3]),(arr[i,0]):(arr[i,0]+arr[i,2])]
Roi = Thresh[(arr[(i-1),1]):(arr[(i-1),1]+arr[(i-1),3]),(arr[(i-1),0]):(arr[(i-1),0]+arr[(i-1),2])]
murg = cv2.resize(Roi,(300,200))
murg1 = cv2.resize(Roi1,(300,200))
hori = np.concatenate((murg,murg1),axis=1)
cv2.imshow("horizon",hori)
#print(hori)
elif i == 2 :
Roi2 = Thresh[(arr[i,1]):(arr[i,1]+arr[i,3]),(arr[i,0]):(arr[i,0]+arr[i,2])]
Roi1 = Thresh[(arr[(i-1),1]):(arr[(i-1),1]+arr[(i-1),3]),(arr[(i-1),0]):(arr[(i-1),0]+arr[(i-1),2])]
Roi = Thresh[(arr[(i-2),1]):(arr[(i-2),1]+arr[(i-2),3]),(arr[(i-2),0]):(arr[(i-2),0]+arr[(i-2),2])]
murg = cv2.resize(Roi,(300,200))
murg1 = cv2.resize(Roi1,(300,200))
murg2 = cv2.resize(Roi2,(300,200))
hori = np.concatenate((murg,murg1,murg2),axis=1)
cv2.imshow("horizon",hori)
#print(hori)
elif i == 3 :
Roi3 = Thresh[(arr[i,1]):(arr[i,1]+arr[i,3]),(arr[i,0]):(arr[i,0]+arr[i,2])]
Roi2 = Thresh[(arr[(i-1),1]):(arr[(i-1),1]+arr[(i-1),3]),(arr[(i-1),0]):(arr[(i-1),0]+arr[(i-1),2])]
Roi1 = Thresh[(arr[(i-2),1]):(arr[(i-2),1]+arr[(i-2),3]),(arr[(i-2),0]):(arr[(i-2),0]+arr[(i-2),2])]
Roi = Thresh[(arr[(i-3),1]):(arr[(i-3),1]+arr[(i-3),3]),(arr[(i-3),0]):(arr[(i-3),0]+arr[(i-3),2])]
murg = cv2.resize(Roi,(300,200))
murg1 = cv2.resize(Roi1,(300,200))
murg2 = cv2.resize(Roi2,(300,200))
murg3 = cv2.resize(Roi3,(300,200))
hori = np.concatenate((murg,murg1,murg2,murg3),axis=1)
cv2.imshow("horizon",hori)
#print(hori)
this section is so that I can print the ROI matrix and to select the ROI
if k == ord("1"):
print(murg)
if k == ord("2"):
print(murg1)
if k == ord ("3"):
print(murg2)
if k == ord("4"):
print(murg3)
cv2.imshow('Frame', Thresh)
if k == ord("e"):
break
if k == ord("s"):
roi = cv2.selectROI("Frame", Thresh, fromCenter=False,showCrosshair=False)
tracker = OBJECT_TRACKING_MACHINE['mosse']()
trackers.add(tracker, Thresh, roi)
#print(boxes,success)
cap.release()
cv2.destroyAllWindows()
when you will run this code you can extract 4 ROI frames which will track your ROI's (I haven't added the precaution for empty matrix so it will give you error if you select more than 4 roi's)
my end goal is to extract those ROI videos for Image processing (this code is not done yet and there's more image processing is going to happen in letter part) **
I am using a fisheye camera for depth image. I had tried with StereoSGBM() method for the depth map which has not given the desired results but now I am trying it with StereoBM() method.
stereoMatcher = cv2.StereoBM() preset = 2 ndisparities = 32 WinSize = 5
if not left.grab() or not right.grab():
print("No more frames")
break
_, leftFrame = left.retrieve()
#leftFrame = cropHorizontal(leftFrame)
leftHeight, leftWidth = leftFrame.shape[:2]
_, rightFrame = right.retrieve()
#rightFrame = cropHorizontal(rightFrame)
rightHeight, rightWidth = rightFrame.shape[:2]
if (leftWidth, leftHeight) != imageSize:
print("Left camera has different size than the calibration data")
break
if (rightWidth, rightHeight) != imageSize:
print("Right camera has different size than the calibration data")
break
fixedLeft = cv2.remap(leftFrame, leftMapX, leftMapY, REMAP_INTERPOLATION)
fixedRight = cv2.remap(rightFrame, rightMapX, rightMapY, REMAP_INTERPOLATION)
grayLeft = cv2.cvtColor(fixedLeft, cv2.COLOR_BGR2GRAY)
grayRight = cv2.cvtColor(fixedRight, cv2.COLOR_BGR2GRAY)
As compute a function of StereoBM() also needs a third argument as disparity. I calculated it by using
grayLeft = cv2.cvtColor(fixedLeft, cv2.COLOR_BGR2GRAY)
grayRight = cv2.cvtColor(fixedRight, cv2.COLOR_BGR2GRAY)
disparity = stereoMatcher.compute(grayLeft, grayRight)
depth = stereoMatcher.compute(grayLeft, grayRight)
#cv2.imshow('left', fixedLeft)
#cv2.imshow('right', fixedRight)
cv2.imshow('depth', depth / DEPTH_VISUALIZATION_SCALE)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
left.release() right.release() cv2.destroyAllWindows()
But it is throwing an error which is:
Traceback (most recent call last):
File "fisheye_depth.py", line 104, in <module>
disparity = stereoMatcher.compute(grayLeft, grayRight)
TypeError: Incorrect type of self (must be 'StereoMatcher' or its derivative)
If anyone has any idea about it that what mistake I am doing. Let me know what mistake I am doing taking out the disparity for compute operator().
I'm trying to get the following script to work. I found it from this website but I keep getting the error :
contours, hierarchy = cv2.findContours(thresh1,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)
ValueError: too many values to unpack (expected 2)
I have little experience in Python and OpenCV so if anyone here can help that would be much appreciated!
I'm running Mac OS X 10.13.4, OpenCV 3.4.1, Python 3.6.5
and
Here is the script I'm trying to get work :
import cv2
import numpy as np
cap = cv2.VideoCapture(0)
while( cap.isOpened() ) :
ret,img = cap.read()
gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
blur = cv2.GaussianBlur(gray,(5,5),0)
ret,thresh1 = cv2.threshold(blur,70,255,cv2.THRESH_BINARY_INV+cv2.THRESH_OTSU)
contours, hierarchy = cv2.findContours(thresh1,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)
drawing = np.zeros(img.shape,np.uint8)
max_area=0
for i in range(len(contours)):
cnt=contours[i]
area = cv2.contourArea(cnt)
if(area>max_area):
max_area=area
ci=i
cnt=contours[ci]
hull = cv2.convexHull(cnt)
moments = cv2.moments(cnt)
if moments['m00']!=0:
cx = int(moments['m10']/moments['m00']) # cx = M10/M00
cy = int(moments['m01']/moments['m00']) # cy = M01/M00
centr=(cx,cy)
cv2.circle(img,centr,5,[0,0,255],2)
cv2.drawContours(drawing,[cnt],0,(0,255,0),2)
cv2.drawContours(drawing,[hull],0,(0,0,255),2)
cnt = cv2.approxPolyDP(cnt,0.01*cv2.arcLength(cnt,True),True)
hull = cv2.convexHull(cnt,returnPoints = False)
if(1):
defects = cv2.convexityDefects(cnt,hull)
mind=0
maxd=0
for i in range(defects.shape[0]):
s,e,f,d = defects[i,0]
start = tuple(cnt[s][0])
end = tuple(cnt[e][0])
far = tuple(cnt[f][0])
dist = cv2.pointPolygonTest(cnt,centr,True)
cv2.line(img,start,end,[0,255,0],2)
cv2.circle(img,far,5,[0,0,255],-1)
print(i)
i=0
cv2.imshow('output',drawing)
cv2.imshow('input',img)
k = cv2.waitKey(10)
if k == 27:
break
Thank you so much in advanced!
The function cv2.fincContours has changed since OpenCV 3.x , and back in OpenCV 4.x . Try this:
## Find contours
if cv2.__version__.startswith("3."):
_, contours, _ = cv2.findContours(thresh1,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)
else:
contours, _ = cv2.findContours(thresh1,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)
Related:
How to use `cv2.findContours` in different OpenCV versions?
I'm using following code to add watermark to animated GIF images. My problem is that all GIF frames except the first one have incorrect colors in result. Would you know how to fix the color of frames? Thank you.
def add_watermark(in_file, watermark_file, watermark_position, watermark_ratio, out_file, quality=85):
img = Image.open(in_file)
watermark_layer = Image.new('RGBA', img.size, (0,0,0,0))
watermark_img = Image.open(watermark_file).convert('RGBA')
watermark_img.thumbnail((img.size[0]/watermark_ratio, 1000), Image.ANTIALIAS)
alpha = watermark_img.split()[3]
alpha = ImageEnhance.Brightness(alpha).enhance(0.95)
watermark_img.putalpha(alpha)
watermark_layer.paste(watermark_img, count_watermark_position(img, watermark_img, watermark_position))
frames = images2gif.readGifFromPIL(img, False)
frames_out = []
for frame in frames:
frames_out.append(Image.composite(watermark_layer, frame, watermark_layer))
images2gif.writeGif(out_file, frames_out, duration=0.5)
To complete example, i provide also code of helper function:
def count_watermark_position(img, watermark, position):
if position == 'right_bottom':
return img.size[0] - watermark.size[0], img.size[1] - watermark.size[1]
if position == 'center':
return (img.size[0] - watermark.size[0])/2, (img.size[1] - watermark.size[1])/2
if position == 'left_bottom':
return 0, img.size[1] - watermark.size[1]
if position == 'left_top':
return 0, 0
if position == 'right_top':
return img.size[0] - watermark.size[0], 0
raise AttributeError('Invalid position')
Source code of images2gif I 've used - I modified it a little bit to make it work with pillow. See comment at the begining of source code.
I've gotten OpenCV working with Python and I can even detect a face through my webcam. What I really want to do though, is see movement and find the point in the middle of the blob of movement. The camshift sample is close to what I want, but I don't want to have to select which portion of the video to track. Bonus points for being able to predict the next frame.
Here's the code I have currently:
#!/usr/bin/env python
import cv
def is_rect_nonzero(r):
(_,_,w,h) = r
return (w > 0) and (h > 0)
class CamShiftDemo:
def __init__(self):
self.capture = cv.CaptureFromCAM(0)
cv.NamedWindow( "CamShiftDemo", 1 )
self.storage = cv.CreateMemStorage(0)
self.cascade = cv.Load("/usr/local/share/opencv/haarcascades/haarcascade_mcs_upperbody.xml")
self.last_rect = ((0, 0), (0, 0))
def run(self):
hist = cv.CreateHist([180], cv.CV_HIST_ARRAY, [(0,180)], 1 )
backproject_mode = False
i = 0
while True:
i = (i + 1) % 12
frame = cv.QueryFrame( self.capture )
if i == 0:
found = cv.HaarDetectObjects(frame, self.cascade, self.storage, 1.2, 2, 0, (20, 20))
for p in found:
# print p
self.last_rect = (p[0][0], p[0][1]), (p[0][2], p[0][3])
print self.last_rect
cv.Rectangle( frame, self.last_rect[0], self.last_rect[1], cv.CV_RGB(255,0,0), 3, cv.CV_AA, 0 )
cv.ShowImage( "CamShiftDemo", frame )
c = cv.WaitKey(7) % 0x100
if c == 27:
break
if __name__=="__main__":
demo = CamShiftDemo()
demo.run()
Found a solution at How do I track motion using OpenCV in Python?