#print n_p.shape
if n_p.size != 0:
v = np.concatenate((v, n_p))
f += 1
else:
f += 1
v = np.concatenate((v, pts))
print "No of circles detected : {0}.".format(num_of_circles)
plt.figure("Detecting Circles and Marking")
#Original Image
plt.subplot(3, 2, 1)
plt.title('Original Image')
plt.imshow(u.convertPlot(im))
plt.xticks([]), plt.yticks([])
#GrayScale Image
plt.subplot(3, 2, 2)
plt.gray()
plt.title('Grayscale Image')
plt.imshow(gray)
plt.xticks([]), plt.yticks([])
#Gaussian Blurred Image
plt.subplot(3, 2, 3)
plt.gray()
plt.title('Blurred Image')
plt.imshow(fgray)
plt.xticks([]), plt.yticks([])
gray = np.float32(gray)
dst = cv2.cornerHarris(gray,2,3,0.04)
#result dilated to mark the corners,i.e we dilate the corner pixels
kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE,(20,20))
dst = cv2.dilate(dst,None)
#thresholding for optimal value, it depends on the image
ret,dst = cv2.threshold(dst,0.045*dst.max(),255,0)
dst = np.uint8(dst)
#finding the centroid
ret,labels,stats,centroids = cv2.connectedComponentsWithStats(dst)
#define the criteria to stop and refine the corners
criteria = (cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 100, 0.001)
corners = cv2.cornerSubPix(gray,np.float32(centroids),(5,5),(-1,-1),criteria)
#Now we draw these
res = np.hstack((centroids,corners))
res = np.int0(res)
img[res[:,1],res[:,0]] = [0,0,255]
img[res[:,3],res[:,2]] = [0,255,0]
"""
plt.figure("Harris Corner Detection");
plt.imshow(utils.convertPlot(img))
plt.show()
print image.shape
#when resizing an image the aspect ratio must be maintained
r = 100.0 / image.shape[1]
#now if the x is fixed as 100
# y becomes r*y
dim = (100, int(image.shape[0] * r))
#performig the resize
resized = cv2.resize(image, dim, interpolation=cv2.INTER_AREA)
subplot(141)
title('Original Image')
imshow(convertPlot(image))
subplot(142)
title('Resized Image (Maintains Aspect ratio)')
imshow(convertPlot(resized))
#rotating the image
(h, w) = image.shape[:2]
center = (w / 2, h / 2)
#rotating the image by 18- degress
M = cv2.getRotationMatrix2D(center, 180, 1.0)
rotated = cv2.warpAffine(image, M, (w, h))
subplot(143)
title('Rotated Image 180')
imshow(convertPlot(rotated))
'snaps/I8.jpg',
'snaps/I9.jpg',
'snaps/I10.jpg'
]
image = cv2.imread(IMAGE_PATHS[0])
marker = find_marker(image)
focalLength = (marker[1][0]*KNOWN_DISTANCE)/KNOWN_WIDTH
#looping over the images,to get the distance
for imagePath in IMAGE_PATHS:
image = cv2.imread(imagePath)
marker = find_marker(image)
inches = distance_from_camera(KNOWN_WIDTH, focalLength, marker[1][0])
box = np.int0(cv2.boxPoints(marker))
cv2.drawContours(image, [box], -1, (0, 255, 0), 2)
cv2.putText(image, "%.2fft" % (inches / 12),
(image.shape[1] - 200, image.shape[0] - 20), cv2.FONT_HERSHEY_SIMPLEX,
2.0, (0, 255, 0), 3)
plt.figure("Distance from Camera")
plt.imshow(utils.convertPlot(image))
plt.show()
cv2.waitKey(2000)
import numpy as np
if __name__ == '__main__':
face_cascade = cv2.CascadeClassifier('/home/riaz/opencv-3.0.0/data/haarcascades/haarcascade_frontalface_default.xml')
eye_cascade = cv2.CascadeClassifier('/home/riaz/opencv-3.0.0/data/haarcascades/haarcascade_eye.xml')
img = cv2.imread('book.jpg')
copy = img.copy()
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
faces = face_cascade.detectMultiScale(gray, 1.3, 5)
for (x, y, w, h) in faces:
cv2.rectangle(img, (x, y), (x+w, y+h), (255, 0, 0), 2)
roi_gray = gray[y:y+h, x:x+w]
roi_color = img[y:y+h, x:x+w]
eyes = eye_cascade.detectMultiScale(roi_gray)
for (ex, ey, ew, eh) in eyes:
cv2.rectangle(roi_color, (ex, ey), (ex+ew, ey+eh), (0, 255, 0), 2)
plt.figure("Face Detection")
plt.subplot(1, 2, 1)
plt.imshow(u.convertPlot(copy))
plt.xticks([]), plt.yticks([])
plt.subplot(1, 2, 2)
plt.imshow(u.convertPlot(img))
plt.xticks([]), plt.yticks([])
plt.show()
for c in cnts:
#approximate the contour
peri = cv2.arcLength(c, True)
#if the approx poly has 4 points,we can conclude the contour to be the screen, since this also happens to be the largest contour
approx = cv2.approxPolyDP(c, 0.02 * peri, True)
if len(approx) == 4:
scrnCnt = approx
break
#OpenCV comes with a draw contour functionality to mark these contours
#we use that the library function to mark the contours
print scrnCnt
cv2.drawContours(copy, [scrnCnt], -1, (0, 255, 0), 3)
plt.figure("Recognizing rectangle")
#plt.gray()
plt.subplot(1, 3, 1)
plt.imshow(utils.convertPlot(im))
plt.subplot(1, 3, 2)
plt.imshow(edged)
plt.subplot(1, 3, 3)
plt.imshow(utils.convertPlot(copy))
plt.show()
rect = (
50, 50, 450, 290
) #this rect can be made interactive later, by selecting using the mouse
"""
Parameter:
1.image
2. mask (0 filled image mask)
3. rect (of interest)
4. background model
5. foreground model
6. interationCount - to have a better segmentation
7. mode: since we use a rectangle box: cv2.GC_INIT_WITH_RECT
"""
img = utils.convertPlot(img)
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]
newmask = cv2.imread('grabCut_manual.png', 0)
#whatever is marked white is foreground ,mask = 1
#whatever is marked black is background , mask = 0
mask[newmask == 0] = 0 #background
mask[newmask == 255] = 1 #foreground
mask, bgdModel, fgdModel = cv2.grabCut(img, mask, None, bgdModel, fgdModel, 5,
cv2.GC_INIT_WITH_MASK)
mask = np.where((mask == 2) | (mask == 0), 0, 1).astype('uint8')
#print n_p.shape
if n_p.size != 0:
v = np.concatenate((v, n_p))
f += 1
else:
f += 1
v = np.concatenate((v, pts))
print "No of circles detected : {0}.".format(num_of_circles)
plt.figure("Detecting Circles and Marking")
#Original Image
plt.subplot(3, 2, 1)
plt.title('Original Image')
plt.imshow(u.convertPlot(im))
plt.xticks([]), plt.yticks([])
#GrayScale Image
plt.subplot(3, 2, 2)
plt.gray()
plt.title('Grayscale Image')
plt.imshow(gray)
plt.xticks([]), plt.yticks([])
#Gaussian Blurred Image
plt.subplot(3, 2, 3)
plt.gray()
plt.title('Blurred Image')
plt.imshow(fgray)
plt.xticks([]), plt.yticks([])
fgdModel= np.zeros((1,65),np.float64)
rect = (50,50,450,290 )#this rect can be made interactive later, by selecting using the mouse
"""
Parameter:
1.image
2. mask (0 filled image mask)
3. rect (of interest)
4. background model
5. foreground model
6. interationCount - to have a better segmentation
7. mode: since we use a rectangle box: cv2.GC_INIT_WITH_RECT
"""
img = utils.convertPlot(img)
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]
newmask = cv2.imread('grabCut_manual.png',0)
#whatever is marked white is foreground ,mask = 1
#whatever is marked black is background , mask = 0
mask[newmask == 0] = 0 #background
mask[newmask == 255] = 1 #foreground
mask, bgdModel, fgdModel = cv2.grabCut(img,mask,None,bgdModel,fgdModel,5,cv2.GC_INIT_WITH_MASK)
mask = np.where((mask==2)|(mask==0),0,1).astype('uint8')
print image.shape
#when resizing an image the aspect ratio must be maintained
r = 100.0/image.shape[1]
#now if the x is fixed as 100
# y becomes r*y
dim = (100, int(image.shape[0]*r))
#performig the resize
resized = cv2.resize(image,dim,interpolation=cv2.INTER_AREA)
subplot(141)
title('Original Image')
imshow(convertPlot(image))
subplot(142)
title('Resized Image (Maintains Aspect ratio)')
imshow(convertPlot(resized))
#rotating the image
(h,w) = image.shape[:2]
center = (w/2,h/2)
#rotating the image by 18- degress
M = cv2.getRotationMatrix2D(center,180,1.0)
rotated=cv2.warpAffine(image,M,(w,h))
subplot(143)
title('Rotated Image 180')
imshow(convertPlot(rotated))
#Using naive Harris Corner Detection with Kernel size 20
import cv2
import numpy as np
from matplotlib import pyplot as plt
import utils
filename = "chessBoard.jpg"
img = cv2.imread(filename)
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
gray = np.float32(gray)
dst = cv2.cornerHarris(gray, 2, 3, 0.04)
#result dilated to mark the corners,i.e we dilate the corner pixels
kernel = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (20, 20))
dst = cv2.dilate(dst, kernel)
#thresholding for optimal value, it depends on the image
img[dst > 0.045 * dst.max()] = [255, 0, 0
] #put blue color for pixels, > 0.01*dst.max()
plt.figure("Harris Corner Detection")
plt.imshow(utils.convertPlot(img))
plt.show()
for c in cnts:
#approximate the contour
peri = cv2.arcLength(c, True)
#if the approx poly has 4 points,we can conclude the contour to be the screen, since this also happens to be the largest contour
approx = cv2.approxPolyDP(c, 0.02 * peri, True)
if len(approx) == 4:
scrnCnt = approx
break
#OpenCV comes with a draw contour functionality to mark these contours
#we use that the library function to mark the contours
print scrnCnt
cv2.drawContours(copy, [scrnCnt], -1, (0, 255, 0), 3)
plt.figure("Recognizing rectangle")
#plt.gray()
plt.subplot(1, 3, 1)
plt.imshow(utils.convertPlot(im))
plt.subplot(1, 3, 2)
plt.imshow(edged)
plt.subplot(1, 3, 3)
plt.imshow(utils.convertPlot(copy))
plt.show()
#This part does blob detection,
sbd = cv2.SimpleBlobDetector_create()
#we store the detected blobs or keypoints in the image
kpts = sbd.detect(smooth)
#now we plot these keypoints in the image
blob = mark_blobs(smooth, kpts, (0, 255, 0), 10)
ret, normal = cv2.threshold(smooth, 100, 255, cv2.THRESH_BINARY)
inv = 255 - normal
"""
plt.figure("Detect book in an image")
plt.subplot(2, 3, 1)
plt.imshow(utils.convertPlot(im))
plt.xticks([]), plt.yticks([])
plt.subplot(2, 3, 2)
plt.gray()
plt.imshow(smooth)
plt.xticks([]), plt.yticks([])
plt.subplot(2, 3, 3)
plt.imshow(thres)
plt.xticks([]), plt.yticks([])
plt.subplot(2, 3, 4)
plt.gray()
plt.imshow(canny)
