Python OpenCV contour only seems to trace the edges of the image, not the object in the middle-CodePudding

So i've been using OpenCV on python 3.6 to approximate a shape as a polygon, and plot the polygonal approximation on the binary image. Below, I have my code and the end-result picture.

As you can see here, the green line (the polygonal contour), only traces the edge of the image, so it only counts 4 sides (b/c the picture is a rectangle), completely disregarding the object in the middle.

No clue what's going on. My code worked well with a similar image.

img = cv2.imread(img) 
black = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) 
# Apply threshold (just in case gray is not binary image).

# apply morphology close
kernel = cv2.getStructuringElement(cv2.MORPH_RECT, (3, 3)) ;
morph = cv2.morphologyEx(black, cv2.MORPH_CLOSE, kernel) ;

# get contours and keep largest
contours = cv2.findContours(morph, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE) ;
contours = contours[0] if len(contours) == 2 else contours[1]  ;
big_contour = max(contours, key=cv2.contourArea) ;

# draw contour
contour = img.copy() ;
cv2.drawContours(contour, [big_contour], 0, (0,0,255), 1) ; # get number of vertices (sides)
peri = cv2.arcLength(big_contour, True)
approx = cv2.approxPolyDP(big_contour, 0.001 * peri, True)

protrusions = (len(approx) - 3)/3
print('number of sides:',len(approx))
print("number of protrusions: ",protrusions)

cv2.drawContours(contour, [approx], -1, (0, 255, 0), 1)
cv2.drawContours(contour, approx, -1, (255, 0, 0), 2)



# save results
cv2.imwrite("quadrilateral_edges.jpg", edges) ;
cv2.imwrite("quadrilateral_morphology.jpg", morph) ;
cv2.imwrite("quadrilateral_contour.jpg", contour) ;

CodePudding user response：

So i changed one thing instead of using cv2.cvtColor I used a cv2.Canny which gave a good result.

Change the line black = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

to black = cv2.Canny(img, 90, 130)

here is an screenshotof the output