Ask Your Question
0

Unable to find and draw biggest contour

asked 2019-12-24 17:06:33 -0600

hernancrespo gravatar image

I want to find and draw biggest contour on a frame. i can find the contours. then i add these three lines to find biggest contour.

 c = max(contours, key = cv2.contourArea)
 x,y,w,h = cv2.boundingRect(c)
 cv2.rectangle(fg,(x,y),(x+w,y+h),(0,255,0),2

and i see an error message:

max() arg is an empty sequence

but when i run len(contours) i see a value as a result, contour is not empty. here is my code:

import cv2
import numpy as np
#capture.release()
cv2.destroyAllWindows()
capture = cv2.VideoCapture(1)
panel = np.zeros([100, 700], np.uint8)
cv2.namedWindow('panel')
def nothing(x):
    pass

cv2.createTrackbar('L - h', 'panel', 0, 179, nothing) #
cv2.createTrackbar('U - h', 'panel', 179, 179, nothing)

cv2.createTrackbar('L - s', 'panel', 0, 255, nothing)
cv2.createTrackbar('U - s', 'panel', 255, 255, nothing)

cv2.createTrackbar('L - v', 'panel', 0, 255, nothing)
cv2.createTrackbar('U - v', 'panel', 255, 255, nothing)

cv2.createTrackbar('S ROWS', 'panel', 0, 480, nothing) #start 
cv2.createTrackbar('E ROWS', 'panel', 480, 480, nothing) # end
cv2.createTrackbar('S COL', 'panel', 0, 640, nothing)
cv2.createTrackbar('E COL', 'panel', 640, 640, nothing)

while(True):

    ret, frame = capture.read()

    s_r = cv2.getTrackbarPos('S ROWS', 'panel')
    e_r = cv2.getTrackbarPos('E ROWS', 'panel')
    s_c = cv2.getTrackbarPos('S COL', 'panel')
    e_c = cv2.getTrackbarPos('E COL', 'panel')
    #print(frame.shape) = 480 , 640, 3 
    roi = frame[s_r: e_r, s_c: e_c]
    roi =  cv2.GaussianBlur(roi, (5, 5), 0) #sigma = 0
    hsv = cv2.cvtColor( roi, cv2.COLOR_RGB2HSV)    

    l_h = cv2.getTrackbarPos('L - h', 'panel')
    u_h = cv2.getTrackbarPos('U - h', 'panel')
    l_s = cv2.getTrackbarPos('L - s', 'panel')
    u_s = cv2.getTrackbarPos('U - s', 'panel')
    l_v = cv2.getTrackbarPos('L - v', 'panel')
    u_v = cv2.getTrackbarPos('U - v', 'panel')
    lower_green = np.array([l_h, l_s, l_v])
    upper_green = np.array([u_h, u_s, u_v])


    mask = cv2.inRange(hsv, lower_green, upper_green)
    mask_inv = cv2.bitwise_not(mask)

    bg = cv2.bitwise_and( roi,  roi, mask=mask)
    fg = cv2.bitwise_and( roi,  roi, mask=mask_inv)
    gray = cv2.cvtColor(fg, cv2.COLOR_BGR2GRAY)
    contours, hierarchy = cv2.findContours(gray,cv2.RETR_TREE, cv2.CHAIN_APPROX_NONE)

    for contour in contours:
        area = cv2.contourArea(contour)

        if area>20000:
            cv2.drawContours(fg, contours, -1, (0, 255, 0), 3)

    c = max(contours, key = cv2.contourArea)

    x,y,w,h = cv2.boundingRect(c)
    # draw the book contour (in green)
    cv2.rectangle(fg,(x,y),(x+w,y+h),(0,255,0),2)

    cv2.imshow('bg', bg)
    cv2.imshow('fg', fg)
    cv2.imshow('panel', panel)

    if cv2.waitKey(30) == 27: #siradaki frame'e gecmeden once 30 ms bekle
        break

capture.release()
cv2.destroyAllWindows()
edit retag flag offensive close merge delete

Comments

Note that every frame might not have a contour, so you need to check for that anyway and handle those frames. The for loop doesn't care if contours is empty, so max() is the first function in trouble.

mvuori gravatar imagemvuori ( 2019-12-25 02:42:05 -0600 )edit

1 answer

Sort by ยป oldest newest most voted
0

answered 2019-12-26 08:18:52 -0600

supra56 gravatar image

updated 2019-12-26 08:21:42 -0600

WOOwww. That cool! I Solved your problem. The error you getting at:

 c = max(contours, key = cv2.contourArea)

To fix problem: Put inside if/else condition block:

    if area>20000:
        cv2.drawContours(fg, contours, -1, (0, 255, 0), 3)

c = max(contours, key = cv2.contourArea)

x,y,w,h = cv2.boundingRect(c)
# draw the book contour (in green)
cv2.rectangle(fg,(x,y),(x+w,y+h),(0,255,0),2)

cv2.imshow('bg', bg)
cv2.imshow('fg', fg)
cv2.imshow('panel', panel)

to:

    if area>20000:
        cv2.drawContours(fg, contours, -1, (0, 255, 0), 3)

        c = max(contours, key = cv2.contourArea)

        x,y,w,h = cv2.boundingRect(c)
        # draw the book contour (in green)
        cv2.rectangle(fg,(x,y),(x+w,y+h),(0,255,0),2)

cv2.imshow('bg', bg)
cv2.imshow('fg', fg)
cv2.imshow('panel', panel)

Also change this:

capture = cv2.VideoCapture(1)

to:

capture = cv2.VideoCapture(0)
edit flag offensive delete link more

Comments

@hernancrespo. Merry Xmas and Happy New Years. Enjoyed your code.

supra56 gravatar imagesupra56 ( 2019-12-26 08:36:42 -0600 )edit

Question Tools

1 follower

Stats

Asked: 2019-12-24 17:06:33 -0600

Seen: 9,625 times

Last updated: Dec 26 '19