Image-Processing-Final/seg.py at main · deependra227/Image-Processing-Final · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
# -*- coding: utf-8 -*-
"""180227_Assign4 Q6

Automatically generated by Colaboratory.

Original file is located at
    https://colab.research.google.com/drive/1HqwFn35rgs2aTHM-Z9t0wvIR4sZRt27P
"""

import numpy as np
import cv2
from scipy import signal, interpolate
from google.colab.patches import cv2_imshow

def flatter_image(img):
    flat_image = img.flatten()
    edgemin = np.amin(flat_image)
    edgemax = np.amax(flat_image)
    edgedelta = edgemax - edgemin
    return flat_image, edgemin, edgemax,edgedelta
def derived(sigmaspatial, sigmarange, samplespatial, samplerange):

  samplespatial = sigmaspatial if (samplespatial is None) else samplespatial
  samplerange = sigmarange if (samplerange is None) else samplerange
  derivedspatial = sigmaspatial / samplespatial
  derivedrange = sigmarange / samplerange
  return samplespatial, samplerange, derivedspatial, derivedrange
def sample_dimension(width,height,edgedelta,samplespatial,samplerange,xypadding,zpadding):
  sw = int(np.round((width - 1) / samplespatial) + 1 + 2 * xypadding)
  sh = int(np.round((height - 1) / samplespatial) + 1 + 2 * xypadding)
  sd = int(np.round(edgedelta / samplerange) + 1 + 2 * zpadding)
  return sw, sh, sd
def get_dim(image,xgrid,ygrid,edgemin,samplespatial,xypadding,zpadding,samplerange,sw, sh, sd):
  dimx = np.around(xgrid / samplespatial) + xypadding
  dimy = np.around(ygrid / samplespatial) + xypadding
  dimz = np.around((image - edgemin) / samplerange) + zpadding
  flatx,flaty,flatz = flat_them(dimx,dimy,dimz)
  dim = flatz + flaty * sd + flatx * sw * sd
  dim = np.array(dim, dtype=int)
  return dimx,dimy,dimz,dim
def flat_them(dimx,dimy,dimz):
  flatx = dimx.flatten()
  flaty = dimy.flatten()
  flatz = dimz.flatten()
  return flatx,flaty,flatz
def rt_BF(image, sigmaspatial, sigmarange, samplespatial=None, samplerange=None):

    height = image.shape[0]
    width = image.shape[1]

    samplespatial, samplerange, derivedspatial, derivedrange = derived(sigmaspatial, sigmarange, samplespatial, samplerange)

    flat_image, edgemin, edgemax,edgedelta = flatter_image(image)

    flat_image = image.flatten()

    xypadding = np.round(2 * derivedspatial + 1)
    zpadding = np.round(2 * derivedrange + 1)

    sw, sh, sd = sample_dimension(width,height,edgedelta,samplespatial,samplerange,xypadding,zpadding)

    df = np.zeros(sh * sw * sd)

    (ygrid, xgrid) = np.meshgrid(range(width), range(height))

    dimx,dimy,dimz,dim= get_dim(image,xgrid,ygrid,edgemin,samplespatial,xypadding,zpadding,samplerange,sw, sh, sd)

    df[dim] = flat_image

    data = df.reshape(sh, sw, sd)
    weights = np.array(data, dtype=bool)

    kerneldim = derivedspatial * 2 + 1
    kerneldep = 2 * derivedrange * 2 + 1


    (gridx, gridy, gridz) = np.meshgrid(range(int(kerneldim)), range(int(kerneldim)), range(int(kerneldep)))

    halfkerneldim = np.round(kerneldim / 2)
    halfkerneldep = np.round(kerneldep / 2)
    gridx -= int(halfkerneldim)
    gridy -= int(halfkerneldim)
    gridz -= int(halfkerneldep)

    gridsqr = ((gridx * gridx + gridy * gridy) / (derivedspatial * derivedspatial)) + ((gridz * gridz) / (derivedrange * derivedrange))
    kernel = np.exp(-0.5 * gridsqr)

    blurweights = signal.fftconvolve(weights, kernel, mode='same')
    blurweights = np.where(blurweights == 0, -2, blurweights)

    blurdata = signal.fftconvolve(data, kernel, mode='same')

    normalblurdata = blurdata / blurweights
    normalblurdata = np.where(blurweights < -1, 0, normalblurdata)

    (ygrid, xgrid) = np.meshgrid(range(width), range(height))

    dimx = (xgrid / samplespatial) + xypadding
    dimy = (ygrid / samplespatial) + xypadding
    dimz = (image - edgemin) / samplerange + zpadding

    out_img = interpolate.interpn((range(normalblurdata.shape[0]), range(normalblurdata.shape[1]),
                               range(normalblurdata.shape[2])), normalblurdata, (dimx, dimy, dimz))
    return out_img

img = cv2.imread("iitk.jpg")

lab = cv2.cvtColor(img,cv2.COLOR_BGR2LAB)
l, a, b = cv2.split(lab)
l=l/100.0
l = l.astype("float32")
result = rt_BF(l, 75.0, 0.2)
result= result*100.0
result = np.uint8(result)
result = cv2.merge((result,a,b))
result = cv2.cvtColor(result,cv2.COLOR_LAB2BGR)

import numpy as np
import cv2

def my_SEG(img):
  # img = result.copy()
  img = cv2.cvtColor(img,cv2.COLOR_BGR2RGB)
  img2d = img.reshape((-1,3))
  img2d = np.float32(img2d)

  _, label, _ = cv2.kmeans(img2d, 2, None, (cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 10, 0.5), 5, cv2.KMEANS_PP_CENTERS)

  label_reshaped = label.reshape((img.shape[:2]))
  label_reshaped = label_reshaped.astype('uint8')
  _, thresh1 = cv2.threshold(label_reshaped, 0, 255, cv2.THRESH_BINARY_INV)

  b = cv2.GaussianBlur(thresh1,(13,13),5)

  _, out = cv2.threshold(b, 200, 255, cv2.THRESH_BINARY)

  return out


img_seg = my_SEG(result.copy())
cv2_imshow(img_seg)
cv2.imwrite('fordrone.jpg', img_seg)