CODEWITHSUNDEEP
pythonopencv
Sandip Mukherjee
Sandip Mukherjee

Reputation: 3

How to extract image inside bounding box which was drawn manually

I have many images which have bounding boxes all drawn in the same color and are all rectangles. How can I extract the image inside the bounding box? There are similar questions but none of those solutions work properly. Can anybody help please? The below image is an example. I would like to use python Opencv to do this.

Image with rectangle regions to extract

EDIT------------------------------------------------------------------

This code works perfectly for all the images. It's based on answer from sipho

img = cv2.imread(img_url)
hsv_min = np.array([0, 250, 100],np.uint8)
hsv_max = np.array([10, 255, 255],np.uint8)

hsv_img = cv2.cvtColor(img,cv2.COLOR_BGR2HSV)

frame_threshed = cv2.inRange(hsv_img, hsv_min, hsv_max)

# Perform morphology
se = np.ones((1,1), dtype='uint8')
image_close = cv2.morphologyEx(frame_threshed, cv2.MORPH_CLOSE, se)
cv2.imshow(image_close)
cv2.waitkey(0)    

# detect contours on the morphed image
ret,thresh = cv2.threshold(image_close,127,255,0)
contours, hierarchy = cv2.findContours(thresh,cv2.RETR_TREE,cv2.CHAIN_APPROX_SIMPLE)

areaArray = []
for i, c in enumerate(contours):
    area = cv2.contourArea(c)
    areaArray.append(area)

# Sort countours based on area
sorteddata = sorted(zip(areaArray, contours), key=lambda x: x[0], reverse=True)

# find the nth largest contour [n-1][1], in this case 2
largestcontour = sorteddata[0][1]

# get the bounding rectangle of the contour
x, y, w, h = cv2.boundingRect(largestcontour)

cropped_img = img[y+3:y+h-3,x+3:x+w-3]
cv2.imshow('cropped', cropped_img)
cv2.waitkey(0)

Upvotes: 0

Views: 2858

Answers (1)

Sipho
Sipho

Reputation: 188

You could try the following:

  1. Convert image to HSV
  2. Threshold image with colors in a certain range
  3. Detect contours in the thresholded image
  4. Filter contours using hierachy to only extract inner contours
  5. Extract ROIs based on the bounding rectangles of the contours

If you are still completely stumped, view this example code

Upvotes: 1

Related Questions