ValueError: not enough values to unpack (expected 3, got 2)

[ypj95]

❓ Questions and Help

So I followed the jupyter notebook tutorial and putted the code sequences into one file .
When I compile with python I get this error result:

Traceback (most recent call last):
File "demo.py", line 42, in
predictions = coco_demo.run_on_opencv_image(image)
File "/home/nero/Documents/workspace/github/maskrcnn-benchmark/demo/predictor.py", line 179, in run_on_opencv_image
result = self.overlay_mask(result, top_predictions)
File "/home/nero/Documents/workspace/github/maskrcnn-benchmark/demo/predictor.py", line 292, in overlay_mask
thresh, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE
ValueError: not enough values to unpack (expected 3, got 2)

Seems that the last saved value in predictor.py in overlay_mask() can't be found.. not sure what to do.

Running on Ubuntu 18.04
Cuda 10 & 9.1

[kishorepv]

The problem is in this line in maskrcnn-benchmark/demo/predictor.py:

_, contours, hierarchy = cv2.findContours(thresh, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE) # fails with error "not enough values to unpack (expected 3, got 2)"

Change it to:

contours, hierarchy = cv2.findContours(thresh, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)

Restart the kernel and run. Your problem should be fixed

[ypj95]

That's what I have tried and it worked too. Just wondering why expecting a 3rd parameter at all.

Cheers for help!

[fmassa]

Hi,

This is a problem with your OpenCV version. The current implementation works fine for OpenCV 3.4.3. If your OpenCV version is < that, then the fix proposed by @kirkshoop should fix it.

[ypj95]

@fmassa Hello,
okey, seems that I am using OpenCV 4.0.0.21. Either way @kirkshoop solution fixed the problem.

[muzammil360]

I also got the above error and fix mentioned above works. However, I am using opencv version '4.0.0'

[mythrireddy]

please tell me solution for this problem

[muzammil360]

@mythrireddy, a simple solution to this problem is to follow the advice of @kishorepv above.

[DeepakSuryaS]

@kishorepv your solution worked. Thanks man.

[pathfinderconsultancyhr123]

@kishorepv reddy saab , problem solved thanks man, can i have your mail id?

[Piyush142]

ret,contours, hierarchy = cv2.findContours(bordered, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)

Change it to
contours, hierarchy = cv2.findContours(bordered, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)

[AnnAfloatX]

How do I get the modified img in this case then? Does it mean that the source img is modified?
(using verions 4.1.0 btw)

[muzammil360]

Does it mean that the source img is modified?

@AnnMLin, since you are using 4.1.0, the source image is not modified in this case. Take a look at documentation here.

How do I get the modified img in this case then?

What do you mean by modified image? If you want to modify your input image (in whatever way you choose), you can read each counter from contours and apply it as your wish.

I hope this helps.

[snehitvaddi]

@AnnMLin @barrelsrider
Since you are using opencv > 4, just use
contours, hierarchy = cv2.findContours(img, cv2.RETR_CCOMP, cv2.CHAIN_APPROX_NONE)
instead of
image,contours, hierarchy = cv2.findContours(img, cv2.RETR_CCOMP, cv2.CHAIN_APPROX_NONE)
Though without using image value it worked fine. It is drawing contours just with values of contours and hierarchy returned. It doesn't require any image or ret value to be declared.

[Rahesh31]

The problem is in this line in maskrcnn-benchmark/demo/predictor.py:
_, contours, hierarchy = cv2.findContours(thresh, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE) # fails with error "not enough values to unpack (expected 3, got 2)"
Change it to:
contours, hierarchy = cv2.findContours(thresh, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
Restart the kernel and run. Your problem should be fixed

I have this problem what should I do now Im new to this

[muzammil360]

@rahesh, you need to remove the "underscore" from relevent file. The solution has been explained in detail above.

[snehitvaddi]

@Rahesh31 bro, refer above discussions you will have clear idea.

[Falcatrua]

from opencv 4.0 on,

contours, find_h = cv2.findContours(im, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE)
contour_img = np.zeros(im.shape)
cv2.drawContours(contour_img, contours, -1, tuple([255]*im.shape[-1]), 1)

contour_img will have only the contours drawn in white

[Rahesh31]

Tq so much 😊i got it

…

On Thu 5 Sep, 2019, 6:17 PM Falcatrua, ***@***.***> wrote: from opencv 4.0 on, contours, find_h = cv2.findContours(im, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE) contour_img = np.zeros(im.shape) cv2.drawContours(contour_img, contours, -1, tuple([255]*im.shape[-1]), 1) contour_img will have only the contours drawn in white — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#339?email_source=notifications&email_token=AM7K3BH5CI67ZN3YB5NGRIDQID5WZA5CNFSM4GPURHUKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD5666AQ#issuecomment-528346882>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AM7K3BDTCMQDQPOFDJJFIV3QID5WZANCNFSM4GPURHUA> .

[jholar]

ValueError: too many values to unpack (expected 2)

Please help me to solve this error
Thanks in advance
i am using opencv 3.4.5. version

[botcs]

just modify L74 to contours, hierarchy, _, but really, I'm curious where did you get the idea that a windows OCR script error should be issued here? :D

[jholar]

Hey sorry for posting on wrong post but really thanks for replying the issue i am getting is solved now!!!!!!

[itsayushsingh]

I had the same issue. Working fine with the above fix. Thanks :D

[mamarasulovs]

The problem is in this line in maskrcnn-benchmark/demo/predictor.py:

_, contours, hierarchy = cv2.findContours(thresh, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE) # fails with error "not enough values to unpack (expected 3, got 2)"

Change it to:

contours, hierarchy = cv2.findContours(thresh, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)

Restart the kernel and run. Your problem should be fixed

thanks dear. its working

[tnavadiya]

Only this given photo works. If I tried some other photos which have garden in background then it doesn't worked.

[tonnidewan]

Tq so much 😊i got it
…
On Thu 5 Sep, 2019, 6:17 PM Falcatrua, @.***> wrote: from opencv 4.0 on, contours, find_h = cv2.findContours(im, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_NONE) contour_img = np.zeros(im.shape) cv2.drawContours(contour_img, contours, -1, tuple([255]*im.shape[-1]), 1) contour_img will have only the contours drawn in white — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#339?email_source=notifications&email_token=AM7K3BH5CI67ZN3YB5NGRIDQID5WZA5CNFSM4GPURHUKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD5666AQ#issuecomment-528346882>, or mute the thread https://github.com/notifications/unsubscribe-auth/AM7K3BDTCMQDQPOFDJJFIV3QIur5WZANCNFSM4GPURHUA .

Scalar value for argument 'color' is longer than 4 . this error is showed when used your code

[tejaswinivjadhav]

@fmassa it fixed the problem .thank you

[sowbhagya09]

i changed it to " contours, hierarchy = cv2.findContours(thresh, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)"
then it is getting like this :

in
63 closing = cv2.cvtColor(closing,cv2.COLOR_BGR2GRAY)
64
---> 65 contours, hierarchy = cv2.findContours(thresh, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE) # find contours with simple approximation cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE
66
67 closing = cv2.cvtColor(closing,cv2.COLOR_GRAY2RGB)

error: OpenCV(4.1.2) C:\projects\opencv-python\opencv\modules\imgproc\src\contours.cpp:197: error: (-210:Unsupported format or combination of formats) [Start]FindContours supports only CV_8UC1 images when mode != CV_RETR_FLOODFILL otherwise supports CV_32SC1 images only in function 'cvStartFindContours_Impl'

[towhidrahman]

[towhidrahman]

help me anyone

[cu2uyash]

Hi,

This is a problem with your OpenCV version. The current implementation works fine for OpenCV 3.4.3. If your OpenCV version is < that, then the fix proposed by @kirkshoop should fix it.

@fmassa that works thank you

[av04121997]

---> 34 image, contours, hierarchy= cv2.findContours(thresh, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE )
35
36

ValueError: not enough values to unpack (expected 3, got 2)

[nithinreddyy]

just modify L74 to contours, hierarchy, _, but really, I'm curious where did you get the idea that a windows OCR script error should be issued here? :D

It's not working for me can you please look into it.

[nithinreddyy]

Hey sorry for posting on the

wrong post but really thanks for replying the issue i am getting is solved now!!!!!!

How did you get that? I'm getting the same error again

[Tam678]

The solution above fixed my problem,thanks a lot!

[shubhamnagalwade]

@kishorepv Your solution work for me. Thank you.

[nesrine20mezni]

please i need help ??

[Leon-Brant]

你好，

这是您的 OpenCV 版本的问题。当前的实现适用于 OpenCV 3.4.3。如果您的 OpenCV 版本是 < 那，那么@kirkshoop提出的修复应该修复它。
Thank you for sharing this information. It is more important to know the cause of the error than to fix it

[zhhvjhtgrfedfgrhjuurgtrf]

import cv2
import numpy as np
import math

def distance(x1, y1, x2, y2):
"""
Calculate distance between two points
"""
dist = math.sqrt(math.fabs(x2-x1)**2 + math.fabs(y2-y1)**2)
return dist

def find_color1(frame):
"""
Filter "frame" for HSV bounds for color1 (inplace, modifies frame) & return coordinates of the object with that color
"""
hsv_frame = cv2.cvtColor(frame, cv2.COLOR_BGR2HSV)
hsv_lowerbound = np.array([102, 152, 0]) #replace THIS LINE w/ your hsv lowerb
hsv_upperbound = np.array([118, 255, 255])#replace THIS LINE w/ your hsv upperb
mask = cv2.inRange(hsv_frame, hsv_lowerbound, hsv_upperbound)
res = cv2.bitwise_and(frame, frame, mask=mask) #filter inplace
, cnts, = cv2.findContours(mask, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
if len(cnts) > 0:
maxcontour = max(cnts, key=cv2.contourArea)

    #Find center of the contour 
    M = cv2.moments(maxcontour)
    if M['m00'] > 0 and cv2.contourArea(maxcontour) > 1000:
        cx = int(M['m10']/M['m00'])
        cy = int(M['m01']/M['m00'])
        return (cx, cy), True
    else:
        return (700, 700), False #faraway point
else:
    return (700, 700), False #faraway point

def find_color2(frame):
"""
Filter "frame" for HSV bounds for color1 (inplace, modifies frame) & return coordinates of the object with that color
"""
hsv_frame = cv2.cvtColor(frame, cv2.COLOR_BGR2HSV)
hsv_lowerbound = np.array([40, 83, 0])#replace THIS LINE w/ your hsv lowerb
hsv_upperbound = np.array([101, 255, 255])#replace THIS LINE w/ your hsv upperb
mask = cv2.inRange(hsv_frame, hsv_lowerbound, hsv_upperbound)
res = cv2.bitwise_and(frame, frame, mask=mask)
,cnts, = cv2.findContours(mask, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
if len(cnts) > 0:
maxcontour = max(cnts, key=cv2.contourArea)

    #Find center of the contour 
    M = cv2.moments(maxcontour)
    if M['m00'] > 0 and cv2.contourArea(maxcontour) > 2000:
        cx = int(M['m10']/M['m00'])
        cy = int(M['m01']/M['m00'])
        return (cx, cy), True #True
    else:
        return (700, 700), True #faraway point
else:
    return (700, 700), True #faraway point

cap = cv2.VideoCapture(0)

while(1):
_, orig_frame = cap.read()

if orig_frame is None:
    break


#we'll be inplace modifying frames, so save a copy
copy_frame = orig_frame.copy() 
(color1_x, color1_y), found_color1 = find_color1(copy_frame)
(color2_x, color2_y), found_color2 = find_color2(copy_frame)

#draw circles around these objects
cv2.circle(copy_frame, (color1_x, color1_y), 20, (255, 0, 0), -1)
cv2.circle(copy_frame, (color2_x, color2_y), 20, (0, 128, 255), -1)

if found_color1 and found_color2:
    #trig stuff to get the line
    hypotenuse = distance(color1_x, color1_x, color2_x, color2_y)
    horizontal = distance(color1_x, color1_y, color2_x, color1_y)
    vertical = distance(color2_x, color2_y, color2_x, color1_y)
    angle = np.arcsin(vertical/hypotenuse)*180.0/math.pi

    #draw all 3 lines
    cv2.line(copy_frame, (color1_x, color1_y), (color2_x, color2_y), (0, 0, 255), 2)
    cv2.line(copy_frame, (color1_x, color1_y), (color2_x, color1_y), (0, 0, 255), 2)
    cv2.line(copy_frame, (color2_x, color2_y), (color2_x, color1_y), (0, 0, 255), 2)

    #put angle text (allow for calculations upto 180 degrees)
    angle_text = ""
    if color2_y < color1_y and color2_x > color1_x:
        angle_text = str(int(angle))
    elif color2_y < color1_y and color2_x < color1_x:
        angle_text = str(int(180 - angle))
    elif color2_y > color1_y and color2_x < color1_x:
        angle_text = str(int(180 + angle))
    elif color2_y > color1_y and color2_x > color1_x:
        angle_text = str(int(360 - angle))
    
    #CHANGE FONT HERE
    cv2.putText(copy_frame, angle_text, (color1_x-30, color1_y), cv2.FONT_HERSHEY_COMPLEX, 1, (0, 128, 229), 2)

cv2.imshow('AngleCalc', copy_frame)
cv2.waitKey(5) 

cap.release()
cv2.destroyAllWindows()

error

Traceback (most recent call last):
File "C:/Users/Besitzer/Real-Time-and-Static-Angles-Calculation-main/Real-Time-and-Static-Angles-Calculation-main/Angle_Findings Taha/color_based.py", line 72, in
(color1_x, color1_y), found_color1 = find_color1(copy_frame)
File "C:/Users/Besitzer/Real-Time-and-Static-Angles-Calculation-main/Real-Time-and-Static-Angles-Calculation-main/Angle_Findings Taha/color_based.py", line 21, in find_color1
, cnts, = cv2.findContours(mask, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)
ValueError: not enough values to unpack (expected 3, got 2)
[ WARN:1@4.056] global D:\a\opencv-python\opencv-python\opencv\modules\videoio\src\cap_msmf.cpp (539) `anonymous-namespace'::SourceReaderCB::~SourceReaderCB terminating async callback

Process finished with exit code 1