CODEWITHSUNDEEP
pythonopencvconvexity-defects
TechieBoy101
TechieBoy101

Reputation: 304

Opencv:convexityDefects on largest contour gives error

I am getting this error :

OpenCV Error: Assertion failed (hpoints > 0) in cv::convexityDefects, file C:\projects\opencv-python\opencv\modules\imgproc\src\convhull.cpp, line 284
Traceback (most recent call last):
  File "E:/PycharmProjects/ComputerVisionAgain/Image Segmentation/hand_blk/main.py", line 12, in <module>
    hull_defects=cv2.convexityDefects(sorted_cnts[0],hull)
cv2.error: C:\projects\opencv-python\opencv\modules\imgproc\src\convhull.cpp:284: error: (-215) hpoints > 0 in function cv::convexityDefects

when I try to get the convexityDefects of the largest contour of the image. This is the code I am using:

import cv2
import numpy as np

img=cv2.imread('blk_hand.jpg')
gray=cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)

ret,thresh=cv2.threshold(gray,100,255,cv2.THRESH_BINARY)

_,contours,h=cv2.findContours(thresh,cv2.RETR_LIST,cv2.CHAIN_APPROX_NONE)
sorted_cnts=sorted(contours,key=cv2.contourArea,reverse=True)
hull=cv2.convexHull(sorted_cnts[0])
hull_defects=cv2.convexityDefects(sorted_cnts[0],hull)
cv2.drawContours(img,[hull],-1,(0,0,255),3)

cv2.drawContours(img,sorted_cnts[0],-1,(0,255,0),3)

cv2.imshow('img',img)
cv2.imshow('thresh',thresh)
cv2.waitKey(0)

This is the original image

This is threshed image

This is convex hulled image on largest contour

Upvotes: 1

Views: 3225

Answers (1)

hannarud
hannarud

Reputation: 131

cv2.convexHull returns convex hull as a set of points by default (returnPoints argument is responsible for this, defaulted to True, docs here). But the cv2.convexityDefects function, according to docs, expects 2nd argument to be indices of the contour points that make the hull.

So just change

hull=cv2.convexHull(sorted_cnts[0])

to

hull=cv2.convexHull(sorted_cnts[0], returnPoints=False)

so hull will contain indices of original sorted_cnts[0] contour that make the convex hull.

BTW, in this case you still can obtain the hull as a set of points by doing sorted_cnts[0][hull].

Upvotes: 6

Related Questions