Crop Image from all sides after edge detection

Question

I am new to OpenCV. I want to extract the main object from an image. So, I have applied Canny on the image to get the edges around the main object and got the following output :

Here is the code to get this using OpenCV in Python:

img = cv2.imread(file)
cv2.imshow("orig", img)
cv2.waitKey(0)
img = cv2.blur(img,(2,2))
gray_seg = cv2.Canny(img, 0, 50)

Now, I want to have the below image as the final output after getting only the main object in the image :

I want to do it in an optimized manner because I have to process more that 2.5 million images. Can anyone help me with this?

I would presume your canny returns you a matrix of magnitudes. Just set a threshold and find the left-most, right-most, top-most, and bottom-most pixel crossing the threshold and cut along vertical and horizontal lines going through found points. — Tomasz Kaminski, Nov 30 '15 at 12:15
Do you require this to be done with OpenCV, or would an [ImageJ](http://imagej.net/) script do as well? — Jan Eglinger, Nov 30 '15 at 12:31
@TomaszKaminski Can you please explain how the image will be cut along those horizontal and vertical lines? — meemee, Nov 30 '15 at 12:38
Just get the `boundingRect` of the edge mask, and slice the image accordingly. — Miki, Nov 30 '15 at 12:43

score 6 · Answer 1 · edited Aug 06 '20 at 14:07

As you have found clean canny edge, then to crop the rectangle region, you should calculate the rectangle boundary.

The steps:

The result:

To calculate the canny region boundary, you can just find no-zero points, then get the min-max coords. Easy to realize using NumPy. Then crop using slice-op.

#!/usr/bin/python3
# 2018.01.20 15:18:36 CST

#!/usr/bin/python3
# 2018.01.20 15:18:36 CST

import cv2
import numpy as np
#img = cv2.imread("test.png")
img = cv2.imread("img02.png")
blurred = cv2.blur(img, (3,3))
canny = cv2.Canny(blurred, 50, 200)

## find the non-zero min-max coords of canny
pts = np.argwhere(canny>0)
y1,x1 = pts.min(axis=0)
y2,x2 = pts.max(axis=0)

## crop the region
cropped = img[y1:y2, x1:x2]
cv2.imwrite("cropped.png", cropped)

tagged = cv2.rectangle(img.copy(), (x1,y1), (x2,y2), (0,255,0), 3, cv2.LINE_AA)
cv2.imshow("tagged", tagged)
cv2.waitKey()

Of course, do boundingRect on the points works too.

score 1 · Answer 2 · edited May 23 '17 at 10:27

1

The rect function should provide the functionality you need. An example of how to use it can be seen below.

cv::Mat image(img);
cv::Rect myROI(posX, posY, sizeX, sizeY);   
cv::Mat croppedImage = image(myROI);

This is writing in c++ but should be able to find a python equivalent.

The link below my provide more info How to crop a CvMat in OpenCV?

edited May 23 '17 at 10:27

Community

1
1

answered Nov 30 '15 at 18:05

Eni

115
8

Please note that you will need to convert the image to a matrix to use this function – Eni Nov 30 '15 at 18:07

Crop Image from all sides after edge detection

2 Answers2

Linked