Revision history [back]

When making a program to count seeds, the contrast is poor on the image: And a Backlight isn't possible.

Tried the following WaterShed Algorithms:

• https://stackoverflow.com/questions/25789278/coffee-beans-separation-algorithm
• https://codegolf.stackexchange.com/q/40831/71194
• https://www.pyimagesearch.com/2015/11/02/watershed-opencv/

Without success, due to the poor contrast and shadow between the Seeds.

To improve remove the shadow between the seeds, GrabCut was tried.

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

img = cv2.imread("C:\\image\\img.bmp")

mask = np.zeros(img.shape[:2],np.uint8)

bgdModel = np.zeros((1,65),np.float64)
fgdModel = np.zeros((1,65),np.float64)

rect = (0,0,1023,767)
cv2.grabCut(img,mask,rect,bgdModel,fgdModel,5,cv2.GC_INIT_WITH_RECT)

mask2 = np.where((mask==2)|(mask==0),0,1).astype('uint8')
img = img*mask2[:,:,np.newaxis]
cv2.imshow('Imagem', img)
cv2.waitKey(0)
cv2.imwrite('C:\\Users\\Desktop\\teste\\resultado.jpg', img)
plt.imshow(img),plt.colorbar(),plt.show()

And the following result:

Question

It wasn't possible to remove the shadows between the seeds, to improve the performance of the watershed algorithm. Is there a better way to remove the sadhows between objects?

When making a program to count seeds, the contrast is poor on the image: And a Backlight isn't possible.

Tried the following WaterShed Algorithms:

• https://stackoverflow.com/questions/25789278/coffee-beans-separation-algorithm
• https://codegolf.stackexchange.com/q/40831/71194
• https://www.pyimagesearch.com/2015/11/02/watershed-opencv/

Without success, due to the poor contrast and shadow between the Seeds.

To improve remove the shadow between the seeds, GrabCut was tried.

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

img = cv2.imread("C:\\image\\img.bmp")

mask = np.zeros(img.shape[:2],np.uint8)

bgdModel = np.zeros((1,65),np.float64)
fgdModel = np.zeros((1,65),np.float64)

rect = (0,0,1023,767)
cv2.grabCut(img,mask,rect,bgdModel,fgdModel,5,cv2.GC_INIT_WITH_RECT)

mask2 = np.where((mask==2)|(mask==0),0,1).astype('uint8')
img = img*mask2[:,:,np.newaxis]
cv2.imshow('Imagem', img)
cv2.waitKey(0)
cv2.imwrite('C:\\Users\\Desktop\\teste\\resultado.jpg', cv2.imwrite('C:\\Users\\Desktop\\test\\result.bmp', img)
plt.imshow(img),plt.colorbar(),plt.show()

And the following result:

Question

It wasn't possible to remove the shadows between the seeds, to improve the performance of the watershed algorithm. Is there a better way to remove the sadhows between objects?

When making a program to count seeds, the contrast is poor on the image: And a Backlight isn't possible.

Tried the following WaterShed Algorithms:

• https://stackoverflow.com/questions/25789278/coffee-beans-separation-algorithm
• https://codegolf.stackexchange.com/q/40831/71194
• https://www.pyimagesearch.com/2015/11/02/watershed-opencv/

Without success, due to the poor contrast and shadow between the Seeds.

To improve remove the shadow between the seeds, GrabCut was tried.

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

img = cv2.imread("C:\\image\\img.bmp")

mask = np.zeros(img.shape[:2],np.uint8)

bgdModel = np.zeros((1,65),np.float64)
fgdModel = np.zeros((1,65),np.float64)

rect = (0,0,1023,767)
cv2.grabCut(img,mask,rect,bgdModel,fgdModel,5,cv2.GC_INIT_WITH_RECT)

mask2 = np.where((mask==2)|(mask==0),0,1).astype('uint8')
img = img*mask2[:,:,np.newaxis]
cv2.imshow('Imagem', img)
cv2.waitKey(0)
cv2.imwrite('C:\\Users\\Desktop\\test\\result.bmp', img)
plt.imshow(img),plt.colorbar(),plt.show()

And the following result:

Question

It wasn't possible to remove the shadows between the seeds, to improve the performance of the watershed algorithm. Is there a better way to remove the sadhows between objects?

When making a program to count seeds, the contrast is poor on the image: And a Backlight isn't possible.

Tried the following WaterShed Algorithms:

• https://stackoverflow.com/questions/25789278/coffee-beans-separation-algorithm
• https://codegolf.stackexchange.com/q/40831/71194
• https://www.pyimagesearch.com/2015/11/02/watershed-opencv/

Without success, due to the poor contrast and shadow between the Seeds.

To improve remove the shadow between the seeds, GrabCut was tried.

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

img = cv2.imread("C:\\image\\img.bmp")

mask = np.zeros(img.shape[:2],np.uint8)

bgdModel = np.zeros((1,65),np.float64)
fgdModel = np.zeros((1,65),np.float64)

rect = (0,0,1023,767)
cv2.grabCut(img,mask,rect,bgdModel,fgdModel,5,cv2.GC_INIT_WITH_RECT)

mask2 = np.where((mask==2)|(mask==0),0,1).astype('uint8')
img = img*mask2[:,:,np.newaxis]
cv2.imshow('Imagem', img)
cv2.waitKey(0)
cv2.imwrite('C:\\Users\\Desktop\\test\\result.bmp', img)
plt.imshow(img),plt.colorbar(),plt.show()

And the following result:

Question

It wasn't possible to remove the shadows between the seeds, to improve the performance of the watershed algorithm. Is there a better way to remove the sadhows between objects?

When making a program to count seeds, the contrast is poor on the image: And a Backlight isn't possible.

Tried the following WaterShed Algorithms:

• https://stackoverflow.com/questions/25789278/coffee-beans-separation-algorithm
• https://codegolf.stackexchange.com/q/40831/71194
• https://www.pyimagesearch.com/2015/11/02/watershed-opencv/

Without success, due to the poor contrast and shadow between the Seeds.

To improve remove the shadow between the seeds, GrabCut was tried.

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

img = cv2.imread("C:\\image\\img.bmp")

mask = np.zeros(img.shape[:2],np.uint8)

bgdModel = np.zeros((1,65),np.float64)
fgdModel = np.zeros((1,65),np.float64)

rect = (0,0,1023,767)
cv2.grabCut(img,mask,rect,bgdModel,fgdModel,5,cv2.GC_INIT_WITH_RECT)

mask2 = np.where((mask==2)|(mask==0),0,1).astype('uint8')
img = img*mask2[:,:,np.newaxis]
cv2.imshow('Imagem', img)
cv2.waitKey(0)
cv2.imwrite('C:\\Users\\Desktop\\test\\result.bmp', img)
plt.imshow(img),plt.colorbar(),plt.show()

And the following result:

Question

It wasn't possible to remove the shadows between the seeds, to improve the performance of the watershed algorithm. Is there a better way to remove the sadhows between objects?

When making a program to count seeds, the contrast is poor on the image: And a Backlight isn't possible.

Tried the following WaterShed Algorithms:

• https://stackoverflow.com/questions/25789278/coffee-beans-separation-algorithm
• https://codegolf.stackexchange.com/q/40831/71194
• https://www.pyimagesearch.com/2015/11/02/watershed-opencv/

Without success, due to the poor contrast and shadow between the Seeds.

To improve remove the shadow between the seeds, GrabCut was tried.

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

img = cv2.imread("C:\\image\\img.bmp")

mask = np.zeros(img.shape[:2],np.uint8)

bgdModel = np.zeros((1,65),np.float64)
fgdModel = np.zeros((1,65),np.float64)

rect = (0,0,1023,767)
cv2.grabCut(img,mask,rect,bgdModel,fgdModel,5,cv2.GC_INIT_WITH_RECT)

mask2 = np.where((mask==2)|(mask==0),0,1).astype('uint8')
img = img*mask2[:,:,np.newaxis]
cv2.imshow('Imagem', img)
cv2.waitKey(0)
cv2.imwrite('C:\\Users\\Desktop\\test\\result.bmp', img)
plt.imshow(img),plt.colorbar(),plt.show()

And the following result:

EDIT:

With the code:

import numpy as np
import cv2
from matplotlib import pyplot as plt
from scipy.ndimage import label
import imutils


def show_image(name, img, vf = False):
    cv2.imshow(name, img)
    cv2.waitKey(0)
    if vf:
        plt.imshow(img),plt.colorbar(),plt.show()

img = cv2.imread("C:\\image\\seeds.bmp")
img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# show_image('Imagem Inicial', img, True)

# https://stackoverflow.com/questions/44752240/how-to-remove-shadow-from-scanned-images-using-opencv
dilated_img = cv2.dilate(img, np.ones((13,13), np.uint8))
# show_image('Dilated', dilated_img)

bg_img = cv2.medianBlur(dilated_img, 7)
# show_image('Median Blur', bg_img)

diff_img = 255 - cv2.absdiff(img, bg_img)
# show_image('Diff', diff_img)

_, img_bin = cv2.threshold(diff_img, 128, 255, cv2.THRESH_OTSU | cv2.THRESH_BINARY_INV)
show_image('Bin1', img_bin)
cv2.imwrite('C:\\image\\bin.bmp', img_bin)

The following image was processed:

However, the Watershed Algo is still not working.

Question

It wasn't possible to remove the shadows between the seeds, to improve the performance of the watershed algorithm. Is there a better way to remove the sadhows between objects?

When making a program to count seeds, the contrast is poor on the image: And a Backlight isn't possible.

Tried the following WaterShed Algorithms:

• https://stackoverflow.com/questions/25789278/coffee-beans-separation-algorithm
• https://codegolf.stackexchange.com/q/40831/71194
• https://www.pyimagesearch.com/2015/11/02/watershed-opencv/

Without success, due to the poor contrast and shadow between the Seeds.

To improve remove the shadow between the seeds, GrabCut was tried.

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

img = cv2.imread("C:\\image\\img.bmp")

mask = np.zeros(img.shape[:2],np.uint8)

bgdModel = np.zeros((1,65),np.float64)
fgdModel = np.zeros((1,65),np.float64)

rect = (0,0,1023,767)
cv2.grabCut(img,mask,rect,bgdModel,fgdModel,5,cv2.GC_INIT_WITH_RECT)

mask2 = np.where((mask==2)|(mask==0),0,1).astype('uint8')
img = img*mask2[:,:,np.newaxis]
cv2.imshow('Imagem', img)
cv2.waitKey(0)
cv2.imwrite('C:\\Users\\Desktop\\test\\result.bmp', img)
plt.imshow(img),plt.colorbar(),plt.show()

And the following result:

EDIT:

With the code:

import numpy as np
import cv2
from matplotlib import pyplot as plt
from scipy.ndimage import label
import imutils


def show_image(name, img, vf = False):
    cv2.imshow(name, img)
    cv2.waitKey(0)
    if vf:
        plt.imshow(img),plt.colorbar(),plt.show()

img = cv2.imread("C:\\image\\seeds.bmp")
img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# show_image('Imagem Inicial', img, True)

# https://stackoverflow.com/questions/44752240/how-to-remove-shadow-from-scanned-images-using-opencv
dilated_img = cv2.dilate(img, np.ones((13,13), np.uint8))
# show_image('Dilated', dilated_img)

bg_img = cv2.medianBlur(dilated_img, 7)
# show_image('Median Blur', bg_img)

diff_img = 255 - cv2.absdiff(img, bg_img)
# show_image('Diff', diff_img)

_, img_bin = cv2.threshold(diff_img, 128, 255, cv2.THRESH_OTSU | cv2.THRESH_BINARY_INV)
show_image('Bin1', img_bin)
cv2.imwrite('C:\\image\\bin.bmp', img_bin)

The following image was processed:

However, the Watershed Algo or Blob Count is still not working.

Question

It wasn't possible to remove the shadows between the seeds, to improve the performance of the watershed algorithm. Is there a better way to remove the sadhows between objects?

When making a program to count seeds, the contrast is poor on the image: And a Backlight isn't possible.

Tried the following WaterShed Algorithms:

• https://stackoverflow.com/questions/25789278/coffee-beans-separation-algorithm
• https://codegolf.stackexchange.com/q/40831/71194
• https://www.pyimagesearch.com/2015/11/02/watershed-opencv/

Without success, due to the poor contrast and shadow between the Seeds.

To improve remove the shadow between the seeds, GrabCut was tried.

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

img = cv2.imread("C:\\image\\img.bmp")

mask = np.zeros(img.shape[:2],np.uint8)

bgdModel = np.zeros((1,65),np.float64)
fgdModel = np.zeros((1,65),np.float64)

rect = (0,0,1023,767)
cv2.grabCut(img,mask,rect,bgdModel,fgdModel,5,cv2.GC_INIT_WITH_RECT)

mask2 = np.where((mask==2)|(mask==0),0,1).astype('uint8')
img = img*mask2[:,:,np.newaxis]
cv2.imshow('Imagem', img)
cv2.waitKey(0)
cv2.imwrite('C:\\Users\\Desktop\\test\\result.bmp', img)
plt.imshow(img),plt.colorbar(),plt.show()

And the following result:

EDIT:

With the code:

import numpy as np
import cv2
from matplotlib import pyplot as plt
from scipy.ndimage import label
import imutils


def show_image(name, img, vf = False):
    cv2.imshow(name, img)
    cv2.waitKey(0)
    if vf:
        plt.imshow(img),plt.colorbar(),plt.show()

img = cv2.imread("C:\\image\\seeds.bmp")
img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# show_image('Imagem Inicial', img, True)

# https://stackoverflow.com/questions/44752240/how-to-remove-shadow-from-scanned-images-using-opencv
dilated_img = cv2.dilate(img, np.ones((13,13), np.uint8))
# show_image('Dilated', dilated_img)

bg_img = cv2.medianBlur(dilated_img, 7)
# show_image('Median Blur', bg_img)

diff_img = 255 - cv2.absdiff(img, bg_img)
# show_image('Diff', diff_img)

_, img_bin = cv2.threshold(diff_img, 128, 255, cv2.THRESH_OTSU | cv2.THRESH_BINARY_INV)
show_image('Bin1', img_bin)
cv2.imwrite('C:\\image\\bin.bmp', img_bin)

The following image was processed:

However, the Watershed Algo or Blob Count is still not working.

Question

It wasn't possible to remove the shadows between the seeds, to improve the performance of the watershed algorithm. Is there a better way to remove the sadhows between objects?

When making a program to count seeds, the contrast is poor on the image: And a Backlight isn't possible.

Tried the following WaterShed Algorithms:

• https://stackoverflow.com/questions/25789278/coffee-beans-separation-algorithm
• https://codegolf.stackexchange.com/q/40831/71194
• https://www.pyimagesearch.com/2015/11/02/watershed-opencv/

Without success, due to the poor contrast and shadow between the Seeds.

To improve remove the shadow between the seeds, GrabCut was tried.

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

img = cv2.imread("C:\\image\\img.bmp")

mask = np.zeros(img.shape[:2],np.uint8)

bgdModel = np.zeros((1,65),np.float64)
fgdModel = np.zeros((1,65),np.float64)

rect = (0,0,1023,767)
cv2.grabCut(img,mask,rect,bgdModel,fgdModel,5,cv2.GC_INIT_WITH_RECT)

mask2 = np.where((mask==2)|(mask==0),0,1).astype('uint8')
img = img*mask2[:,:,np.newaxis]
cv2.imshow('Imagem', img)
cv2.waitKey(0)
cv2.imwrite('C:\\Users\\Desktop\\test\\result.bmp', img)
plt.imshow(img),plt.colorbar(),plt.show()

And the following result:

EDIT:

With the code:

import numpy as np
import cv2
from matplotlib import pyplot as plt
from scipy.ndimage import label
import imutils


def show_image(name, img, vf = False):
    cv2.imshow(name, img)
    cv2.waitKey(0)
    if vf:
        plt.imshow(img),plt.colorbar(),plt.show()

img = cv2.imread("C:\\image\\seeds.bmp")
img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# show_image('Imagem Inicial', img, True)

# https://stackoverflow.com/questions/44752240/how-to-remove-shadow-from-scanned-images-using-opencv
dilated_img = cv2.dilate(img, np.ones((13,13), np.uint8))
# show_image('Dilated', dilated_img)

bg_img = cv2.medianBlur(dilated_img, 7)
# show_image('Median Blur', bg_img)

diff_img = 255 - cv2.absdiff(img, bg_img)
# show_image('Diff', diff_img)

_, img_bin = cv2.threshold(diff_img, 128, 255, cv2.THRESH_OTSU | cv2.THRESH_BINARY_INV)
show_image('Bin1', img_bin)
cv2.imwrite('C:\\image\\bin.bmp', img_bin)

The following image was processed:

However, the Watershed Algo or Blob Count is still not working.

Question

It wasn't possible to remove the shadows between the seeds, to improve the performance of the watershed algorithm. Is there a better way to remove the sadhows between objects?