最長的線路檢測 - 蟒蛇 - OpenCV的

Question

我需要檢測給定圖像中最長的線路和我的形象將與此類似：

enter image description here

我拔後嘗試過，但在細化圖像獲取PIX因爲它不是直線。

有沒有其他解決方法呢？

感謝 Tejas的

Answer 1

你覺得這個解決方案的？

我在代碼中包含了解釋。總體思路是做一個閾值提取黑色區域，然後查找輪廓，並選出最長的一個。

Dilation確實已經完成了指針的所有工作，但是我留下了一些替代代碼，以便在需要時查找最長的輪廓。

#!/usr/bin/env python 

import sys 
import numpy as np 
import cv2 
from matplotlib import pyplot as plt 


img = cv2.imread('image.jpg',0) 
print img.shape 
h, w = img.shape[:2] 

# Drop top and bottom area of image with black parts. 
img= img[100:h-100, :] 
h, w = img.shape[:2] 

# Threshold image 
ret,th1 = cv2.threshold(img,50,255,cv2.THRESH_BINARY) 

# get rid of thinner lines 
kernel = np.ones((5,5),np.uint8) 
th1 = cv2.dilate(th1,kernel,iterations = 3) 

# Determine contour of all blobs found 
_, contours0, hierarchy = cv2.findContours(th1.copy(), cv2.RETR_TREE, cv2.CHAIN_APPROX_NONE) 
contours = [cv2.approxPolyDP(cnt, 3, True) for cnt in contours0] 

# Draw all contours 
vis = np.zeros((h, w, 3), np.uint8) 
cv2.drawContours(vis, contours, -1, (128,255,255), 3, cv2.LINE_AA) 

# Draw the contour with maximum perimeter (omitting the first contour which is outer boundary of image 
# Not necessary in this case 
vis2 = np.zeros((h, w, 3), np.uint8) 
perimeter=[] 
for cnt in contours[1:]: 
    perimeter.append(cv2.arcLength(cnt,True)) 
print perimeter 
print max(perimeter) 
maxindex= perimeter.index(max(perimeter)) 
print maxindex 

cv2.drawContours(vis2, contours, maxindex +1, (255,0,0), -1) 


# Show all images 
titles = ['Original Image','Threshold','Contours', 'Result'] 
images=[img, th1, vis, vis2] 
for i in xrange(4): 
    plt.subplot(2,2,i+1) 
    plt.imshow(images[i],'gray') 
    plt.title(titles[i]), plt.xticks([]), plt.yticks([]) 
plt.show() 

[編輯]

您可以廣告多一些代碼來確定輪廓的主軸線爲如下一行：

# Determine and draw main axis 
length = 300 
(x,y),(MA,ma),angle = cv2.fitEllipse(cnt) 
print np.pi , angle 
print angle * np.pi/180.0 
print np.cos(angle * np.pi/180.0) 
x2 = int(round(x + length * np.cos((angle-90) * np.pi/180.0))) 
y2 = int(round(y + length * np.sin((angle-90) * np.pi/180.0))) 
cv2.line(vis2, (int(x), int(y)), (x2,y2), (0,255,0),5) 
print x,y,x2,y2