Ask Your Question

Help For OpenCV Smoothing Contours

asked 2019-11-02 10:44:40 -0500

smartyon gravatar image

updated 2019-11-03 05:16:20 -0500

berak gravatar image

Hi, I need the code about detecting shapes and colours of the shapes in the photo and the colors of the shapes. Now I coded in the PyCharm, but I can't get correct result from the my program because image's contours are not flat completely. So it says "Ellipse" for about the some triangles. I tried Median Blur, bilateral blur image smoothing methods. And I can't get correct result. Anyone can help to me? Thanks..

import cv2
import numpy as np


img_filter = cv2.imread("C:\\Users\\yusuf\\PycharmProjects\\OpenCV\\venv\\soru1.jpg")

gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)

_,threshold = cv2.threshold(gray,240,255,cv2.THRESH_BINARY)

contours,_ = cv2.findContours(threshold,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)

for cnt in contours:

epsilon = 0.001*cv2.arcLength(cnt,True)
approx = cv2.approxPolyDP(cnt,epsilon,True)
x = approx.ravel()[0]
y = approx.ravel()[1]

centres = [] for i in range(len(contours)): moments = cv2.moments(contours[i])

centres.append((int(moments['m10']/moments['m00']), int(moments['m01']/moments['m00']))), centres[-1], 3, -1)

gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) _, threshold = cv2.threshold(gray, 240, 255, cv2.THRESH_BINARY)

contours, _ = cv2.findContours(threshold, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE) ucgen_sayaci=0 dortgen_sayaci=0 daire_sayaci=0

for cnt in contours: approx = cv2.approxPolyDP(cnt, 0.01*cv2.arcLength(cnt, True), True) cv2.drawContours(img, [approx], 0, (0), 5) x = approx.ravel()[0] y = approx.ravel()[1] if len(approx) == 3: ucgen_sayaci = ucgen_sayaci + 1 cv2.putText(img, "Triangle", (x, y), font1, 1, (0)) elif len(approx) == 4: dortgen_sayaci = dortgen_sayaci + 1 cv2.putText(img, "Rectangle", (x, y), font1, 1, (0)) else: daire_sayaci = daire_sayaci + 1 cv2.putText(img, "Circle", (x, y), font1, 1, (0))

print("Ucgen Sayisi:",ucgen_sayaci) print("Dortgen Sayisi:",dortgen_sayaci) print("Daire Sayisi:",daire_sayaci) cv2.imwrite('output.png',img)

cv2.imshow("IMG",img) cv2.imshow("IMG1",img_filter)

cv2.waitKey(0) cv2.destroyAllWindows()

edit retag flag offensive close merge delete


As there are alternatives, a sample image would help understanding the challenge. But even blindly, your tresholding looks quite a lot highlight oriented if there are various colors in the contours.

mvuori gravatar imagemvuori ( 2019-11-02 12:12:08 -0500 )edit

You could always try Bezier curves?

sjhalayka gravatar imagesjhalayka ( 2019-11-02 15:09:31 -0500 )edit
LBerger gravatar imageLBerger ( 2019-11-02 15:11:54 -0500 )edit

1 answer

Sort by ยป oldest newest most voted

answered 2019-11-05 00:13:06 -0500

smartyon gravatar image

updated 2019-11-05 00:13:20 -0500

I solved the problem with firstly change color space to hsv,after Threshold finally gaussian and some other blur methods. I will share the code as soon as on github. Then after I will edit this mention.

edit flag offensive delete link more
Login/Signup to Answer

Question Tools

1 follower


Asked: 2019-11-02 10:43:53 -0500

Seen: 659 times

Last updated: Nov 05 '19