You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
584 lines
18 KiB
584 lines
18 KiB
import math
|
|
import random
|
|
import uuid
|
|
|
|
import cv2
|
|
import matplotlib.pyplot as plt
|
|
import numpy as np
|
|
from django.http import HttpResponse
|
|
|
|
PREFIX = 'media/'
|
|
|
|
DEFAULT_FORMAT = '.jpg'
|
|
|
|
SUPPORT_FILE_FORMAT = ['png', 'jpg', 'bmp', 'jpeg', 'jpe', 'dib', 'pbm', 'pgm', 'ppm', 'tiff', 'tif']
|
|
|
|
|
|
def get_image_name():
|
|
return uuid.uuid1().__str__()
|
|
|
|
|
|
def get_image(request):
|
|
file = request.FILES.get('image')
|
|
suffix = file.name[file.name.rfind('.') + 1:].lower()
|
|
if suffix not in SUPPORT_FILE_FORMAT:
|
|
return ''
|
|
filename = get_image_name() + '.' + suffix
|
|
img = open(PREFIX + filename, 'wb+')
|
|
for chunk in file.chunks():
|
|
img.write(chunk)
|
|
img.close()
|
|
return filename
|
|
|
|
|
|
def process_bg():
|
|
img = np.zeros((800, 800, 3), np.uint8)
|
|
return save_image(img)
|
|
|
|
|
|
def process_line(para):
|
|
name, start, end, color, thickness = para['img'], para['start'], para['end'], \
|
|
para.get('color', [255, 255, 255]), para.get('thickness', 20)
|
|
img = cv2.imread(PREFIX + name)
|
|
cv2.line(img, tuple(start), tuple(end), tuple(color), thickness)
|
|
cv2.imwrite(PREFIX + name, img)
|
|
return name
|
|
|
|
|
|
def process_rectangle(para):
|
|
name, start, end, color, thickness = para['img'], para['start'], para['end'], \
|
|
para.get('color', [255, 255, 255]), para.get('thickness', 20)
|
|
img = cv2.imread(PREFIX + name)
|
|
cv2.rectangle(img, tuple(start), tuple(end), tuple(color), thickness)
|
|
cv2.imwrite(PREFIX + name, img)
|
|
return name
|
|
|
|
|
|
def process_circle(para):
|
|
name, center, radius, color, thickness = para['img'], para['center'], para['radius'], \
|
|
para.get('color', [255, 255, 255]), para.get('thickness', 20)
|
|
img = cv2.imread(PREFIX + name)
|
|
cv2.circle(img, tuple(center), radius, tuple(color), thickness)
|
|
cv2.imwrite(PREFIX + name, img)
|
|
return name
|
|
|
|
|
|
def process_ellipse(para):
|
|
name, center, axes, angle, startangle, endangle, color, thickness = para['img'], \
|
|
para['center'], \
|
|
para['axes'], \
|
|
para.get('angle', 0), \
|
|
para.get('startangle', 0), \
|
|
para.get('endangle', 360), \
|
|
para.get('color', [255, 255, 255]), \
|
|
para.get('thickness', 20)
|
|
img = cv2.imread(PREFIX + name)
|
|
cv2.ellipse(img, tuple(center), tuple(axes), angle, startangle, endangle, tuple(color), thickness)
|
|
cv2.imwrite(PREFIX + name, img)
|
|
return name
|
|
|
|
|
|
def process_polylines(para):
|
|
name, point, isclosed, color = para['img'], para['point'], para.get('isclosed', True), \
|
|
para.get('color', [255, 255, 255])
|
|
img = cv2.imread(PREFIX + name)
|
|
pts = np.array(point, np.int32)
|
|
pts = pts.reshape((-1, 1, 2))
|
|
cv2.polylines(img, [pts], isclosed, tuple(color))
|
|
cv2.imwrite(PREFIX + name, img)
|
|
return name
|
|
|
|
|
|
def process_text(para):
|
|
name, text, org, fontsize, color = para['img'], para['text'], para['org'], \
|
|
para.get('fontsize', 14), para.get('color', [255, 255, 255])
|
|
img = cv2.imread(PREFIX + name)
|
|
cv2.putText(img, text, org, fontFace=cv2.FONT_HERSHEY_SCRIPT_COMPLEX, fontScale=fontsize, color=tuple(color))
|
|
cv2.imwrite(PREFIX + name, img)
|
|
return name
|
|
|
|
|
|
def process_filter_rgb(para, data):
|
|
limit = para.get('limit', 100)
|
|
c_range = para.get('range', [90, 230])
|
|
if np.max(data) > limit:
|
|
data = data * (255 / np.max(data))
|
|
tmp = np.zeros_like(data)
|
|
tmp[((data > c_range[0]) & (data <= c_range[1]))] = 255
|
|
return tmp
|
|
|
|
|
|
def base2read(img, has_color):
|
|
img[0], img[1] = img[1], img[0]
|
|
img1 = cv2.imread(PREFIX + img[0], has_color)
|
|
img2 = cv2.imread(PREFIX + img[1], has_color)
|
|
rows, cols = img1.shape[:2]
|
|
img2 = cv2.resize(img2, (cols, rows), interpolation=cv2.INTER_CUBIC)
|
|
return img1, img2
|
|
|
|
|
|
def logic_and(img, has_color, base):
|
|
img1 = None
|
|
img2 = None
|
|
if base == 2:
|
|
img1, img2 = base2read(img, has_color)
|
|
ret = img1 & img2
|
|
return save_image(ret)
|
|
|
|
|
|
def logic_or(img, has_color, base):
|
|
img1 = None
|
|
img2 = None
|
|
if base == 2:
|
|
img1, img2 = base2read(img, has_color)
|
|
ret = img1 | img2
|
|
return save_image(ret)
|
|
|
|
|
|
def logic_not(img, has_color, base):
|
|
if base == 2:
|
|
img[0], img[1] = img[1], img[0]
|
|
image = cv2.imread(PREFIX + img[0], has_color)
|
|
ret = ~image
|
|
return save_image(ret)
|
|
|
|
|
|
def arithmetic_add(img, has_color, base):
|
|
img1 = None
|
|
img2 = None
|
|
if base == 2:
|
|
img1, img2 = base2read(img, has_color)
|
|
ret = cv2.add(img1, img2)
|
|
return save_image(ret)
|
|
|
|
|
|
def base2read_choice(img, has_color, base):
|
|
img1 = cv2.imread(PREFIX + img[0], has_color)
|
|
img2 = cv2.imread(PREFIX + img[1], has_color)
|
|
if base == 1:
|
|
rows, cols = img1.shape[:2]
|
|
img2 = cv2.resize(img2, (cols, rows), interpolation=cv2.INTER_CUBIC)
|
|
elif base == 2:
|
|
rows, cols = img2.shape[:2]
|
|
img1 = cv2.resize(img1, (cols, rows), interpolation=cv2.INTER_CUBIC)
|
|
return img1, img2
|
|
|
|
|
|
def arithmetic_sub(img, has_color, base):
|
|
img1, img2 = base2read_choice(img, has_color, base)
|
|
ret = cv2.subtract(img1, img2)
|
|
return save_image(ret)
|
|
|
|
|
|
def arithmetic_multi(img, has_color, base):
|
|
img1 = None
|
|
img2 = None
|
|
if base == 2:
|
|
img1, img2 = base2read(img, has_color)
|
|
ret = cv2.multiply(img1, img2)
|
|
return save_image(ret)
|
|
|
|
|
|
def arithmetic_div(img, has_color, base):
|
|
img1, img2 = base2read_choice(img, has_color, base)
|
|
ret = cv2.divide(img1, img2)
|
|
return save_image(ret)
|
|
|
|
|
|
def rotate_bound(image, angle):
|
|
(h, w) = image.shape[:2]
|
|
(cX, cY) = (w // 2, h // 2)
|
|
M = cv2.getRotationMatrix2D((cX, cY), -angle, 1.0)
|
|
cos = np.abs(M[0, 0])
|
|
sin = np.abs(M[0, 1])
|
|
nW = int((h * sin) + (w * cos))
|
|
nH = int((h * cos) + (w * sin))
|
|
M[0, 2] += (nW / 2) - cX
|
|
M[1, 2] += (nH / 2) - cY
|
|
return cv2.warpAffine(image, M, (nW, nH))
|
|
|
|
|
|
def hist_cover_gray(img, fileName):
|
|
plt.figure(fileName, figsize=(16, 8))
|
|
hist = cv2.calcHist([img], [0], None, [256], [0, 255])
|
|
plt.plot(hist)
|
|
plt.xlim([0, 255])
|
|
plt.savefig(fileName)
|
|
|
|
|
|
def hist_cover_rgb(img, fileName):
|
|
color = ["r", "g", "b"]
|
|
b, g, r = cv2.split(img)
|
|
img = cv2.merge([r, g, b])
|
|
for index, c in enumerate(color):
|
|
hist = cv2.calcHist([img], [index], None, [256], [0, 255])
|
|
plt.plot(hist, color=c)
|
|
plt.xlim([0, 255])
|
|
plt.savefig(fileName)
|
|
|
|
|
|
def grayHist(img, filename):
|
|
plt.figure(filename, figsize=(16, 8))
|
|
h, w = img.shape[:2]
|
|
pixelSequence = img.reshape(h * w, 1)
|
|
numberBins = 256
|
|
histogram_, bins, patch = plt.hist(pixelSequence, numberBins)
|
|
plt.xlabel("gray label")
|
|
plt.ylabel("number of pixels")
|
|
plt.axis([0, 255, 0, np.max(histogram_)])
|
|
plt.savefig(filename)
|
|
|
|
|
|
def process_LoG(img):
|
|
grayImage = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
|
|
img = cv2.copyMakeBorder(grayImage, 2, 2, 2, 2, borderType=cv2.BORDER_REPLICATE)
|
|
img = cv2.GaussianBlur(img, (3, 3), 0, 0)
|
|
m1 = np.array(
|
|
[[0, 0, -1, 0, 0], [0, -1, -2, -1, 0], [-1, -2, 16, -2, -1], [0, -1, -2, -1, 0], [0, 0, -1, 0, 0]],
|
|
dtype=np.int32)
|
|
image1 = np.zeros(img.shape).astype(np.int32)
|
|
h, w, _ = img.shape
|
|
for i in range(2, h - 2):
|
|
for j in range(2, w - 2):
|
|
image1[i, j] = np.sum(m1 * img[i - 2:i + 3, j - 2:j + 3, 1])
|
|
return cv2.convertScaleAbs(image1)
|
|
|
|
|
|
def process_Canny(img):
|
|
tmp = cv2.GaussianBlur(img, (3, 3), 0)
|
|
gradx = cv2.Sobel(tmp, cv2.CV_16SC1, 1, 0)
|
|
grady = cv2.Sobel(tmp, cv2.CV_16SC1, 0, 1)
|
|
return cv2.Canny(gradx, grady, 50, 150)
|
|
|
|
|
|
def process_enhance(img):
|
|
h, w = img.shape
|
|
gradient = np.zeros((h, w))
|
|
img = img.astype('float')
|
|
for i in range(h - 1):
|
|
for j in range(w - 1):
|
|
gx = abs(img[i + 1, j] - img[i, j])
|
|
gy = abs(img[i, j + 1] - img[i, j])
|
|
gradient[i, j] = gx + gy
|
|
sharp = img + gradient
|
|
sharp = np.where(sharp > 255, 255, sharp)
|
|
sharp = np.where(sharp < 0, 0, sharp)
|
|
gradient = gradient.astype('uint8')
|
|
sharp = sharp.astype('uint8')
|
|
ret = [{"sharp": save_image(sharp), "gradient": save_image(gradient)}]
|
|
return HttpResponse(ret)
|
|
|
|
|
|
def process_lcd_with_p(edges, img):
|
|
linesP = cv2.HoughLinesP(edges, 1, np.pi / 180, 80, 200, 15)
|
|
for i_P in linesP:
|
|
for x1, y1, x2, y2 in i_P:
|
|
cv2.line(img, (x1, y1), (x2, y2), (0, 255, 0), 3)
|
|
return img
|
|
|
|
|
|
def process_lcd_without_p(edges, img):
|
|
lines = cv2.HoughLines(edges, 1, np.pi / 2, 118)
|
|
for i_line in lines:
|
|
for line in i_line:
|
|
rho = line[0]
|
|
theta = line[1]
|
|
if (theta < (np.pi / 4.)) or (theta > (3. * np.pi / 4.0)): # 垂直直线
|
|
pt1 = (int(rho / np.cos(theta)), 0)
|
|
pt2 = (int((rho - img.shape[0] * np.sin(theta)) / np.cos(theta)), img.shape[0])
|
|
cv2.line(img, pt1, pt2, (0, 0, 255))
|
|
else:
|
|
pt1 = (0, int(rho / np.sin(theta)))
|
|
pt2 = (img.shape[1], int((rho - img.shape[1] * np.cos(theta)) / np.sin(theta)))
|
|
cv2.line(img, pt1, pt2, (0, 0, 255), 1)
|
|
return img
|
|
|
|
|
|
def ideal_low_filter(img, D0):
|
|
h, w = img.shape[:2]
|
|
filter_img = np.ones((h, w))
|
|
u = np.fix(h / 2)
|
|
v = np.fix(w / 2)
|
|
for i in range(h):
|
|
for j in range(w):
|
|
d = np.sqrt((i - u) ** 2 + (j - v) ** 2)
|
|
filter_img[i, j] = 0 if d > D0 else 1
|
|
return filter_img
|
|
|
|
|
|
def ideal_high_pass_filter(img, D0):
|
|
h, w = img.shape[:2]
|
|
filter_img = np.zeros((h, w))
|
|
u = np.fix(h / 2)
|
|
v = np.fix(w / 2)
|
|
for i in range(h):
|
|
for j in range(w):
|
|
d = np.sqrt((i - u) ** 2 + (j - v) ** 2)
|
|
filter_img[i, j] = 0 if d < D0 else 1
|
|
return filter_img
|
|
|
|
|
|
def butterworth_low_pass_filter(img, D0, rank):
|
|
h, w = img.shape[:2]
|
|
filter_img = np.zeros((h, w))
|
|
u = np.fix(h / 2)
|
|
v = np.fix(w / 2)
|
|
for i in range(h):
|
|
for j in range(w):
|
|
d = np.sqrt((i - u) ** 2 + (j - v) ** 2)
|
|
filter_img[i, j] = 1 / (1 + ((d / D0) ** (2 * rank)))
|
|
return filter_img
|
|
|
|
|
|
def butterworth_high_pass_filter(img, D0, rank):
|
|
h, w = img.shape[:2]
|
|
filter_img = np.zeros((h, w))
|
|
u = np.fix(h / 2)
|
|
v = np.fix(w / 2)
|
|
for i in range(h):
|
|
for j in range(w):
|
|
d = np.sqrt((i - u) ** 2 + (j - v) ** 2)
|
|
filter_img[i, j] = 1 / (1 + ((D0 / d) ** (2 * rank)))
|
|
return filter_img
|
|
|
|
|
|
def gauss_low_pass_filter(img, d0):
|
|
h, w = img.shape[:2]
|
|
filter_img = np.zeros((h, w))
|
|
u = np.fix(h / 2)
|
|
v = np.fix(w / 2)
|
|
for i in range(h):
|
|
for j in range(w):
|
|
d = np.sqrt((i - u) ** 2 + (j - v) ** 2)
|
|
filter_img[i, j] = np.exp(-0.5 * (d ** 2) / (d0 ** 2))
|
|
return filter_img
|
|
|
|
|
|
def gauss_high_pass_filter(img, d0):
|
|
h, w = img.shape[:2]
|
|
filter_img = np.zeros((h, w))
|
|
u = np.fix(h / 2)
|
|
v = np.fix(w / 2)
|
|
for i in range(h):
|
|
for j in range(w):
|
|
d = np.sqrt((i - u) ** 2 + (j - v) ** 2)
|
|
filter_img[i, j] = 1 - np.exp(-0.5 * (d ** 2) / (d0 ** 2))
|
|
return filter_img
|
|
|
|
|
|
def filter_use_smooth(img, my_filter):
|
|
# 首先进行傅里叶变换
|
|
f = np.fft.fft2(img)
|
|
f_center = np.fft.fftshift(f)
|
|
# 应用滤波器进行反变换
|
|
S = np.multiply(f_center, my_filter) # 频率相乘——l(u,v)*H(u,v)
|
|
f_origin = np.fft.ifftshift(S) # 将低频移动到原来的位置
|
|
f_origin = np.fft.ifft2(f_origin) # 使用ifft2进行傅里叶的逆变换
|
|
f_origin = np.abs(f_origin) # 设置区间
|
|
return f_origin
|
|
|
|
|
|
def filter_use_sharpen(img, my_filter):
|
|
f_origin = filter_use_smooth(img, my_filter)
|
|
f_origin = f_origin / np.max(f_origin.all())
|
|
return f_origin
|
|
|
|
|
|
def process_roberts(img):
|
|
kernelx = np.array([[-1, 0], [0, 1]], dtype=int)
|
|
kernely = np.array([[0, -1], [1, 0]], dtype=int)
|
|
x = cv2.filter2D(img, cv2.CV_16S, kernelx)
|
|
y = cv2.filter2D(img, cv2.CV_16S, kernely)
|
|
# 转uint8
|
|
absX = cv2.convertScaleAbs(x)
|
|
absY = cv2.convertScaleAbs(y)
|
|
return cv2.addWeighted(absX, 0.5, absY, 0.5, 0)
|
|
|
|
|
|
def process_sobel(img):
|
|
x = cv2.Sobel(img, cv2.CV_16S, 1, 0) # 对x求一阶导
|
|
y = cv2.Sobel(img, cv2.CV_16S, 0, 1) # 对y求一阶导
|
|
absX = cv2.convertScaleAbs(x)
|
|
absY = cv2.convertScaleAbs(y)
|
|
return cv2.addWeighted(absX, 0.5, absY, 0.5, 0)
|
|
|
|
|
|
def process_prewitt(img):
|
|
kernelx = np.array([[1, 1, 1], [0, 0, 0], [-1, -1, -1]], dtype=int)
|
|
kernely = np.array([[-1, 0, 1], [-1, 0, 1], [-1, 0, 1]], dtype=int)
|
|
x = cv2.filter2D(img, cv2.CV_16S, kernelx)
|
|
y = cv2.filter2D(img, cv2.CV_16S, kernely)
|
|
# 转uint8
|
|
absX = cv2.convertScaleAbs(x)
|
|
absY = cv2.convertScaleAbs(y)
|
|
return cv2.addWeighted(absX, 0.5, absY, 0.5, 0)
|
|
|
|
|
|
def process_laplacian(img):
|
|
dst = cv2.Laplacian(img, cv2.CV_16S, ksize=3)
|
|
return cv2.convertScaleAbs(dst)
|
|
|
|
|
|
def process_gauss_noise(img):
|
|
img = np.array(img / 255, dtype=float)
|
|
noise_ = np.random.normal(0, 0.1, img.shape)
|
|
img = img + noise_
|
|
img = np.clip(img, 0.0, 1.0)
|
|
return np.uint8(img * 255)
|
|
|
|
|
|
def process_salt_and_peper_noise(img, para):
|
|
range_salt = para.get('salt', 0.2)
|
|
range_peper = para.get('peper', 0.8)
|
|
output = np.zeros(img.shape, np.uint8)
|
|
for i in range(img.shape[0]):
|
|
for j in range(img.shape[1]):
|
|
rdn = random.random()
|
|
if rdn < range_salt:
|
|
output[i][j] = 0
|
|
elif rdn > range_peper:
|
|
output[i][j] = 255
|
|
else:
|
|
output[i][j] = img[i][j]
|
|
return output
|
|
|
|
|
|
def process_geometric_mean(img):
|
|
output = np.zeros(img.shape, np.uint8)
|
|
for i in range(img.shape[0]):
|
|
for j in range(img.shape[1]):
|
|
ji = 1.0
|
|
for m in range(-1, 2):
|
|
if 0 <= j + m < img.shape[1]:
|
|
ji *= img[i][j + m]
|
|
output[i][j] = math.pow(ji, 1 / 3)
|
|
return output
|
|
|
|
|
|
def process_arithmetic_mean(img):
|
|
output = np.zeros(img.shape, np.uint8)
|
|
for i in range(img.shape[0]):
|
|
for j in range(img.shape[1]):
|
|
my_sum = 0
|
|
for m in range(-1, 2):
|
|
for n in range(-1, 2):
|
|
if 0 <= i + m < img.shape[0] and 0 <= j + n < img.shape[1]:
|
|
my_sum += img[i + m][j + n]
|
|
output[i][j] = int(my_sum / 9)
|
|
return output
|
|
|
|
|
|
def process_harmonic_filter(img):
|
|
output = np.zeros(img.shape, np.uint8)
|
|
for i in range(img.shape[0]):
|
|
for j in range(img.shape[1]):
|
|
my_sum = 0
|
|
for m in range(-1, 2):
|
|
for n in range(-1, 2):
|
|
if 0 <= i + m < img.shape[0] and 0 <= j + n < img.shape[1]:
|
|
my_sum += 1 / img[i + m][j + n]
|
|
output[i][j] = int(9 / my_sum)
|
|
return output
|
|
|
|
|
|
def get_middle(array):
|
|
length = len(array)
|
|
for i in range(length):
|
|
for j in range(i + 1, length):
|
|
if array[j] > array[i]:
|
|
temp = array[j]
|
|
array[j] = array[i]
|
|
array[i] = temp
|
|
return array[int(length / 2)]
|
|
|
|
|
|
def get_array_from_struct_ele(img, i, j):
|
|
array = []
|
|
for m in range(-1, 2):
|
|
for n in range(-1, 2):
|
|
if 0 <= i + m < img.shape[0] and 0 <= j + n < img.shape[1]:
|
|
array.append(img[i + m][j + n])
|
|
return array
|
|
|
|
|
|
def process_max_sort(img):
|
|
output = np.zeros(img.shape, np.uint8)
|
|
for i in range(img.shape[0]):
|
|
for j in range(img.shape[1]):
|
|
array = get_array_from_struct_ele(img, i, j)
|
|
array.sort(reverse=True)
|
|
output[i][j] = max(array[0], img[i][j])
|
|
return output
|
|
|
|
|
|
def process_min_sort(img):
|
|
output = np.zeros(img.shape, np.uint8)
|
|
for i in range(img.shape[0]):
|
|
for j in range(img.shape[1]):
|
|
array = get_array_from_struct_ele(img, i, j)
|
|
array.sort()
|
|
output[i][j] = min(array[0], img[i][j])
|
|
return output
|
|
|
|
|
|
def process_median_sort(img):
|
|
output = np.zeros(img.shape, np.uint8)
|
|
for i in range(img.shape[0]):
|
|
for j in range(img.shape[1]):
|
|
array = get_array_from_struct_ele(img, i, j)
|
|
output[i][j] = get_middle(array)
|
|
return output
|
|
|
|
|
|
def process_high_choice(img, limit):
|
|
output = np.zeros(img.shape, np.uint8)
|
|
for i in range(img.shape[0]):
|
|
for j in range(img.shape[1]):
|
|
if img[i][j] > limit:
|
|
output[i][j] = img[i][j]
|
|
else:
|
|
output[i][j] = 0
|
|
return output
|
|
|
|
|
|
def process_low_choice(img, limit):
|
|
output = np.zeros(img.shape, np.uint8)
|
|
for i in range(img.shape[0]):
|
|
for j in range(img.shape[1]):
|
|
if img[i][j] < limit:
|
|
output[i][j] = img[i][j]
|
|
else:
|
|
output[i][j] = 0
|
|
return output
|
|
|
|
|
|
def process_pass_choice(img, my_min, my_max):
|
|
output = np.zeros(img.shape, np.uint8)
|
|
for i in range(img.shape[0]):
|
|
for j in range(img.shape[1]):
|
|
if my_min < img[i][j] < my_max:
|
|
output[i][j] = img[i][j]
|
|
else:
|
|
output[i][j] = 255
|
|
return output
|
|
|
|
|
|
def process_stop_choice(img, my_min, my_max):
|
|
output = np.zeros(img.shape, np.uint8)
|
|
for i in range(img.shape[0]):
|
|
for j in range(img.shape[1]):
|
|
if my_min < img[i][j] < my_max:
|
|
output[i][j] = 0
|
|
else:
|
|
output[i][j] = img[i][j]
|
|
return output
|
|
|
|
|
|
def save_image(img, prefix=None):
|
|
img_name = get_image_name() + DEFAULT_FORMAT
|
|
if prefix is None:
|
|
cv2.imwrite(PREFIX + img_name, img)
|
|
else:
|
|
cv2.imwrite(prefix + img_name, img)
|
|
return img_name
|