I am writing a program to detect crop rows in uav imagery and I have successfully managed to detect crop rows in cropped sections of the image using opencv and python, however the issue I am running into now is how to go about detecting the end of a crop row.
The image is of a field(Cant share it publicly) and I am attempting to detect all the rows in that field, I can detect the rows however currently I am simply draw the line along the row from the top to bottom of the image. Instead I would like draw this line from the start to end of the crop row.
i.e. smething like this
https://i.stack.imgur.com/WUkg6.jpg (Not enough rep to post images yet)
But essentially I need to detect the end of the row so I can draw the line to the just the end of the row.
Any Ideas on how this might be done?
Use the Probabilistic Hough Line Transform:
import cv2
import numpy as np
cap = cv2.VideoCapture(0)
frame = cv2.imread("crop.jpg")
# It converts the BGR color space of image to HSV color space
hsv = cv2.cvtColor(frame, cv2.COLOR_BGR2HSV)
# Threshold of blue in HSV space
lower_green = np.array([21, 115, 0])
upper_green = np.array([179, 255, 255])
# preparing the mask to overlay
mask = cv2.inRange(hsv, lower_green, upper_green)
mask = cv2.blur(mask, (5,5))
ret, mask = cv2.threshold(mask, 128, 255, cv2.THRESH_BINARY)
edges = cv2.Canny(mask,100,200)
# The black region in the mask has the value of 0,
# so when multiplied with original image removes all non-blue regions
result = cv2.bitwise_and(frame, frame, mask = mask)
cdst = cv2.cvtColor(mask, cv2.COLOR_GRAY2BGR)
linesP = cv2.HoughLinesP(edges, 1, np.pi / 180, 50, None, 100, 30)
if linesP is not None:
for i in range(0, len(linesP)):
l = linesP[i][0]
cv2.line(cdst, (l[0], l[1]), (l[2], l[3]), (0,0,255), 3, cv2.LINE_AA)
cv2.imshow("Detected Lines (in red) - Probabilistic Line Transform", cdst)
cv2.imshow('frame', frame)
cv2.imshow('mask', mask)
cv2.imshow('result', result)
cv2.waitKey(-1)
Related
I am trying to remove text from images that has a black border with white fill. Take the image below as an example.
I have tried a few options utilizing opencv and skimage inpaint
import cv2
from skimage.restoration import inpaint
img = cv2.imread('Documents/test_image.png')
mask = cv2.threshold(img, 210, 255, cv2.THRESH_BINARY)[1][:,:,0]
dst = cv2.inpaint(img, mask, 7, cv2.INPAINT_TELEA)
image_result = inpaint.inpaint_biharmonic(img, mask,
multichannel=True)
cv2.imshow('image',img)
cv2.imshow('mask',mask)
cv2.imshow('dst',dst)
cv2.imshow('image_result',image_result)
cv2.waitKey(0)
It seems like the inpainting is just trying to fill with black as that is what it is identifying as being around the areas of interest. What I would like to do is remove the white text and black borders completely, or secondarily try to fill the white with more information from surrounding colors than just the black.
Here is the best solution I could come up with, still open to others with more experience showing me a better way if anyone has an idea.
mask = cv2.threshold(img, 245, 255, cv2.THRESH_BINARY)[1][:,:,0]
new_mask = cv2.dilate(mask, cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (10,10)))
dst = cv2.inpaint(img, new_mask, 7, cv2.INPAINT_TELEA)
Here are two inpainting methods in Python/OpenCV. Note that I use the saturation channel to create the threshold, since white and black have zero saturation, in principle.
Input:
import cv2
import numpy as np
# read input
img = cv2.imread('white_black_text.png')
# convert to hsv and extract saturation
hsv = cv2.cvtColor(img,cv2.COLOR_BGR2HSV)
sat = hsv[:,:,1]
# threshold and invert
thresh = cv2.threshold(sat, 10, 255, cv2.THRESH_BINARY)[1]
thresh = 255 - thresh
# apply morphology dilate
kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (15,15))
thresh = cv2.morphologyEx(thresh, cv2.MORPH_DILATE, kernel)
# do inpainting
result1 = cv2.inpaint(img,thresh,11,cv2.INPAINT_TELEA)
result2 = cv2.inpaint(img,thresh,11,cv2.INPAINT_NS)
# save results
cv2.imwrite('white_black_text_threshold.png', thresh)
cv2.imwrite('white_black_text_inpainted1.png', result1)
cv2.imwrite('white_black_text_inpainted2.png', result1)
# show results
cv2.imshow('thresh',thresh)
cv2.imshow('result1',result1)
cv2.imshow('result2',result2)
cv2.waitKey(0)
cv2.destroyAllWindows()
Threshold and morphology cleaned result:
Result 1 (Telea):
Result 2 (Navier Stokes):
I have a picture were I want to change all white-ish pixels to grey, but only for a certain area of the image. Example picture, I just want to change the picture outside of the red rectangle, without changing the image within the red rectangle:
I already have the general code, which was part of someone elses Stackoverflow question, that changes the colour of every white pixel instead of only just the one outside of an area.
image = cv.imread("meme 2.jpg")
hsv = cv.cvtColor(image, cv.COLOR_BGR2HSV)
# Define lower and uppper limits of what we call "white-ish"
sensitivity = 19
lower_white = np.array([0, 0, 255 - sensitivity])
upper_white = np.array([255, sensitivity, 255])
# Mask image to only select white
mask = cv.inRange(hsv, lower_white, upper_white)
# Change image to grey where we found brown
image[mask > 0] = (170, 170, 170)
cv.imwrite(file, image)
Here is one way to do that in Python/OpenCV.
Read the input
Convert to HSV color space
Threshold on desired color to make a mask
Use the mask to change the color of all corresponding pixels in the image
Draw a new rectangular mask for the region where you do not want to change
Invert the new mask for the region where you do want to change
Apply the new mask to the original image
Apply the inverted new mask to the color changed image
Add the two results together to form the final image
Save the results
Input:
import cv2
import numpy as np
# Read image
image = cv2.imread('4animals.jpg')
# Convert to HSV
hsv = cv2.cvtColor(image, cv2.COLOR_BGR2HSV)
# Define lower and uppper limits of what we call "white-ish"
sensitivity = 19
lower_white = np.array([0, 0, 255 - sensitivity])
upper_white = np.array([255, sensitivity, 255])
# Create mask to only select white
mask = cv2.inRange(hsv, lower_white, upper_white)
# Change image to grey where we found white
image2 = image.copy()
image2[mask > 0] = (170, 170, 170)
# Create new rectangular mask that is white on black background
x,y,w,h = 33,100,430,550
mask2 = np.zeros_like(image)
cv2.rectangle(mask2, (x,y), (x+w,y+h), (255, 255, 255), -1)
# invert mask
mask2_inv = 255 - mask2
# apply mask to image
image_masked = cv2.bitwise_and(image, mask2)
# apply inverted mask to image2
image2_masked = cv2.bitwise_and(image2, mask2_inv)
# add together
result = cv2.add(image_masked, image2_masked)
# save results
cv2.imwrite('4animals_mask.jpg', mask)
cv2.imwrite('4animals_modified.png', image2)
cv2.imwrite('4animals_mask2.jpg', mask2)
cv2.imwrite('4animals_mask2_inv.jpg', mask2_inv)
cv2.imwrite('4animals_masked.jpg', image_masked)
cv2.imwrite('4animals_modified_masked.jpg', image2_masked)
cv2.imwrite('4animals_result.jpg', result)
cv2.imshow('mask', mask)
cv2.imshow('image2', image2)
cv2.imshow('mask2', mask2 )
cv2.imshow('mask2_inv', mask2_inv)
cv2.imshow('image_masked', image_masked)
cv2.imshow('image2_masked', image2_masked)
cv2.imshow('result', result)
cv2.waitKey(0)
cv2.destroyAllWindows()
Color mask:
Rectangle mask:
Inverted rectangle mask:
Color changed image:
Masked input:
Masked color changed image:
Result:
Here is another simpler method in Python/OpenCV. My previous answer was overly complicated.
Read the input
Convert to HSV color space
Create a mask image by color thresholding
Draw a black rectangle on the previous mask for where you do not want to change the color
Apply the new combined mask to the image to change the color in the desired region
Save the result
Input:
import cv2
import numpy as np
# Read image
image = cv2.imread('4animals.jpg')
# Convert to HSV
hsv = cv2.cvtColor(image, cv2.COLOR_BGR2HSV)
# Define lower and uppper limits of what we call "white-ish"
sensitivity = 19
lower_white = np.array([0, 0, 255 - sensitivity])
upper_white = np.array([255, sensitivity, 255])
# Create mask to only select white
mask = cv2.inRange(hsv, lower_white, upper_white)
# Draw new rectangular mask on old mask that is black inside the rectangle and white outside the rectangle
x,y,w,h = 33,100,430,550
mask2 = mask.copy()
cv2.rectangle(mask2, (x,y), (x+w,y+h), 0, -1)
# Change image to grey where we found white for combined mask
result = image.copy()
result[mask2 > 0] = (170, 170, 170)
# save results
cv2.imwrite('4animals_mask.jpg', mask)
cv2.imwrite('4animals_mask2.jpg', mask2)
cv2.imwrite('4animals_result.jpg', result)
cv2.imshow('mask', mask)
cv2.imshow('mask2', mask2 )
cv2.imshow('result', result)
cv2.waitKey(0)
cv2.destroyAllWindows()
Mask from color thresholding:
Modified mask with rectangle drawn over it:
Result:
I have road photos from the racing game.
I want to detect yellow and white lanes.
If I use RGB space,
def select_rgb_white_yellow(image):
# white color mask
lower = np.uint8([123, 116, 116])
upper = np.uint8([186, 172, 160])
white_mask = cv2.inRange(image, lower, upper)
# yellow color mask
lower = np.uint8([134, 121, 100])
upper = np.uint8([206, 155, 87])
yellow_mask = cv2.inRange(image, lower, upper)
# combine the mask
mask = cv2.bitwise_or(white_mask, yellow_mask)
masked = cv2.bitwise_and(image, image, mask = mask)
return masked
I only get white lanes.
So tweaked little bit, using HLS space, by using this code.
def convert_hls(image):
return cv2.cvtColor(image, cv2.COLOR_RGB2HLS)
Then, extracting the yellow and white lane again,
def select_white_yellow(image):
converted = convert_hls(image)
# white color mask
lower = np.uint8([0, 0, 0])
upper = np.uint8([0, 0, 255])
white_mask = cv2.inRange(converted, lower, upper)
# yellow color mask
lower = np.uint8([ 10, 0, 100])
upper = np.uint8([ 40, 255, 255])
yellow_mask = cv2.inRange(converted, lower, upper)
# combine the mask
mask = cv2.bitwise_or(white_mask, yellow_mask)
return cv2.bitwise_and(image, image, mask = mask)
Then it cannot detect white lane anymore. Is there good way to detect both white and yellow lane?
Here are all RGB colour code that I found
Yellow
c2974A (194, 149, 74)
a07444 (160, 116, 68)
b38e55 (179, 142, 85)
867964 (134, 121, 100)
ce9b57 (206, 155, 87)
ce9853 (206, 152, 83)
white
b4a59d (180, 165, 157)
b9a99a (185, 169, 154)
baaca0 (186, 172, 160)
867e79 (134, 126, 121)
7b7474 (123, 116, 116)
827d7c (130, 125, 124)
To extend the comment: this is an implementation of a 2 stage approach. Take some time to look at the intermediate images/masks to fully understand everything that happens.
Cropping to region of interest
You can automate this, but I cheated a little and did it manually. The cropped sky area will rarely ever have road surface, this is an easy solution that suffices (for now) I think. Similarly, I also cut of the HUD boxes on the right hand side, as they have similar gray colors as the road and were interfering. It's more neat to draw black boxes over them, so they are excluded from processing.
Isolating road
Convert the cropped image to HSV an select only gray-ish values. After some noise removal I improved the mask using findContours to draw the convexhull. If performance is an issue, you could maybe skip it by tweaking the close mask step.
Selecting lines
Using the mask you can create an image of only the road surface. You can use this for color separation without having to worry about also selecting surroundings. My result is not perfect, but I assume you have larger versions of the images which will give better results.
Result:
Code:
import cv2
import numpy as np
# load image
image = cv2.imread('pw12b.jpg')
# crop image
h,w = image.shape[:2]
image = image[200:h-20,20:550]
# create hsv
hsv = cv2.cvtColor(image, cv2.COLOR_BGR2HSV)
# set lower and upper color limits
low_val = (0,0,0)
high_val = (179,45,96)
# Threshold the HSV image
mask = cv2.inRange(hsv, low_val,high_val)
# remove noise
mask = cv2.morphologyEx(mask, cv2.MORPH_OPEN, kernel=np.ones((8,8),dtype=np.uint8))
# close mask
mask = cv2.morphologyEx(mask, cv2.MORPH_CLOSE, kernel=np.ones((20,20),dtype=np.uint8))
# improve mask by drawing the convexhull
ret, contours, hierarchy = cv2.findContours(mask, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
for cnt in contours:
hull = cv2.convexHull(cnt)
cv2.drawContours(mask,[hull],0,(255), -1)
# erode mask a bit to migitate mask bleed of convexhull
mask = cv2.morphologyEx(mask, cv2.MORPH_ERODE, kernel=np.ones((5,5),dtype=np.uint8))
# remove this line, used to show intermediate result of masked road
road = cv2.bitwise_and(image, image,mask=mask)
# apply mask to hsv image
road_hsv = cv2.bitwise_and(hsv, hsv,mask=mask)
# set lower and upper color limits
low_val = (0,0,102)
high_val = (179,255,255)
# Threshold the HSV image
mask2 = cv2.inRange(road_hsv, low_val,high_val)
# apply mask to original image
result = cv2.bitwise_and(image, image,mask=mask2)
#show image
cv2.imshow("Result", result)
cv2.imshow("Road", road)
cv2.imshow("Mask", mask)
cv2.imshow("Image", image)
cv2.waitKey(0)
cv2.destroyAllWindows()
I'm using OpenCV to find tabular data within images so that I can use an OCR on it. So far I have been able to find the table in the image, find the columns of the table, then find each cell within each column. It works pretty well, but I'm having an issue with the cell walls getting stuck in my images and I'm unable to remove them reliably.This is one example that I'm having difficulty with.This would be another example.
I have tried several approaches to get these images better. I have been having the most luck with finding the contours.
img2gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
ret, mask = cv2.threshold(img2gray, 180, 255, cv2.THRESH_BINARY)
image_final = cv2.bitwise_and(img2gray, img2gray, mask=mask)
ret, new_img = cv2.threshold(image_final, 180, 255, cv2.THRESH_BINARY_INV)
kernel = cv2.getStructuringElement(cv2.MORPH_CROSS, (5, 3))
# dilate , more the iteration more the dilation
dilated = cv2.dilate(new_img, kernel, iterations=3)
cv2.imwrite('../test_images/find_contours_dilated.png', dilated)
I have been toying with the kernel size and dilation iterations and have found this to be the best configuration.
Another approach I used was with PIL, but it is only really good if the border is uniform around the whole image, which in my cases it is not.
copy = Image.fromarray(img)
try:
bg = Image.new(copy.mode, copy.size, copy.getpixel((0, 0)))
except:
return None
diff = ImageChops.difference(copy, bg)
diff = ImageChops.add(diff, diff, 2.0, -100)
bbox = diff.getbbox()
if bbox:
return np.array(copy.crop(bbox))
There were a few other ideas that I tried, but none got me very far. Any help would be appreciated.
You could try with finding contours and "drawing" them out. Meaning you can draw a border that will be connected with the "walls" mannualy with cv2.rectangle on the borders of your image (that will combine all contours - walls). The biggest two contours will be outer and inner line of your walls and you can draw the contour white to remove the border. Then apply threshold again to remove the rest of the noise. Cheers!
Example:
import cv2
import numpy as np
# Read the image
img = cv2.imread('borders2.png')
# Get image shape
h, w, channels = img.shape
# Draw a rectangle on the border to combine the wall to one contour
cv2.rectangle(img,(0,0),(w,h),(0,0,0),2)
# Convert to grayscale
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# Apply binary threshold
_, threshold = cv2.threshold(gray, 180, 255, cv2.THRESH_BINARY_INV)
# Search for contours and sort them by size
_, contours, hierarchy = cv2.findContours(threshold,cv2.RETR_TREE,cv2.CHAIN_APPROX_NONE)
area = sorted(contours, key=cv2.contourArea, reverse=True)
# Draw it out with white color from biggest to second biggest contour
cv2.drawContours(img, ((contours[0]),(contours[1])), -1, (255,255,255), -1)
# Apply binary threshold again to the new image to remove little noises
_, img = cv2.threshold(img, 180, 255, cv2.THRESH_BINARY)
# Display results
cv2.imshow('img', img)
Result:
I have a picture( Basic X-Y plot image, where the plot line is in blue color and x,y axis are in black color), where in that I need to detect the edges based on the color. I came across below code, where its detecting all the lines by using canny edge detection and hough algorithm. But I need to detect only the blue color line in this image. What can i do in order to detect that?
Below is the code that I used.
import numpy as np
import cv2
img = cv2.imread('xyplot.png')
gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
cv2.imshow('grayimage',gray)
edges = cv2.Canny(gray,50,150,apertureSize = 3)
cv2.imshow('edgesimage',edges)
print img.shape[1]
print img.shape
minLineLength=img.shape[1]-300
lines = cv2.HoughLinesP(image=edges,rho=0.02,theta=np.pi/500,
threshold=10,lines=np.array([]),
minLineLength=minLineLength, maxLineGap=100)
a,b,c = lines.shape
for i in range(a):
cv2.line(img, (lines[i][0][0], lines[i][0][1]), (lines[i][0][2],
lines[i][0] [3]), (0, 0, 255), 3, cv2.LINE_AA)
cv2.imshow('edges', edges)
cv2.imshow('result', img)
cv2.waitKey(0)
cv2.destroyAllWindows()
Apply a (blue) color mask on your image before converting to gray scale. You can do so by defining the lower and upper boundary of all 3 channels and performing BITWISE_AND on the original image. You will have to play around with the values of the channel ranges to make sure only the pixels you want get captured.
lower = np.array([200, 20, 20], dtype = "uint8")
upper = np.array([255, 100, 100], dtype = "uint8")
mask = cv2.inRange(img, lower, upper)
img = cv2.bitwise_and(img, img, mask = mask)
Note: The three channels in the list are B, G and R respectively.