def compare_screenshot_to_base(baseline, diff=100):
"""Calculate the exact difference between two images.
:param string baseline: [required] base screenshot to compare
:param int diff: value of maximum difference
Example::
Compare screenshot to base base_screenshot.jpg
"""
path = _get_screenshot()
current_browser = _get_browser()
if hasattr(current_browser, 'get_screenshot_as_file'):
current_browser.get_screenshot_as_file(path)
else:
current_browser.save_screenshot(path)
img1 = Iopen(path)
img2 = Iopen(baseline)
his1 = img1.histogram()
his2 = img2.histogram()
sqrtdiff = lambda a, b: (a - b) ** 2
rms = sqrt(reduce(add, imap(sqrtdiff, his1, his2)) / len(his1))
logger.info("RMS diff: %s" % rms)
if rms > 0:
idiff = difference(img1, img2)
path = path.replace(".png", ".jpg")
idiff.save(path)
logger.info("diff image: %s" % path)
if rms > diff:
raise AssertionError(
"Image: %s is different from baseline: %s" % (path, baseline))
def show(self, flag='image'):
(_, name) = os.path.split(self.image_list[self.current_image_mark])
if flag == 'image':
path = self.path_image
plotwindow = self.image
elif flag == 'GT':
path = self.path_mask_GT
plotwindow = self.mask_GT
elif flag == 'pred':
path = self.path_mask_pred
plotwindow = self.mask_pred
try:
img = Iopen(os.path.join(path, name))
self.data[flag] = array(img)
if (flag == 'GT' and self.data['pred'] is not None) or (
flag == 'pred' and self.data['GT'] is not None):
matrix = evaluate.Matrix(self.data['pred'], self.data['GT'])
_, miou = matrix.miou()
_, acc = matrix.accuracy()
kappa = matrix.kappa()
self.miou.setText(f'{miou}')
self.accuracy.setText(f'{acc}')
self.kappa.setText(f'{kappa}')
self.progressBar.setValue(self.current_image_mark + 1)
img = img.resize((341, 251))
self.file_path.setText(self.image_list[self.current_image_mark])
plotwindow.setPixmap(img.toqpixmap())
self.check_LastAndNext()
except FileNotFoundError as e:
raise (e)
def compare_screenshot_to_base(baseline, diff=100):
"""Calculate the exact difference between two images.
:param string baseline: [required] base screenshot to compare
:param int diff: value of maximum difference
Example::
Compare screenshot to base base_screenshot.jpg
"""
path = _get_screenshot()
current_browser = _get_browser()
if hasattr(current_browser, 'get_screenshot_as_file'):
current_browser.get_screenshot_as_file(path)
else:
current_browser.save_screenshot(path)
img1 = Iopen(path)
img2 = Iopen(baseline)
his1 = img1.histogram()
his2 = img2.histogram()
sqrtdiff = lambda a, b: (a - b)**2
rms = sqrt(reduce(add, imap(sqrtdiff, his1, his2)) / len(his1))
logger.info("RMS diff: %s" % rms)
if rms > 0:
idiff = difference(img1, img2)
path = path.replace(".png", ".jpg")
idiff.save(path)
logger.info("diff image: %s" % path)
if rms > diff:
raise AssertionError("Image: %s is different from baseline: %s" %
(path, baseline))
def plot_eye_scatter(data: DataFrame, ax_in: Axes,
size: (int, int) = (35, 27)) -> list:
"""Standard formatted plot with eye ward symbol scatter.
Arguments:
data {DataFrame} -- Table with xCoordinate, yCoordinate, time,
smoke and Name
ax_in {Axes} -- Target axes for plot
Keyword Arguments:
size {[type]} -- Image size (default: {(14, 34)})
Returns:
list -- Returns list of generated plot entities.
"""
# Map
plot_map(ax_in)
# Add labels
#data = label_smoke_name_time(data, number=True)
data['label'] = ["{}.".format(x + 1) for x in range(data.shape[0])]
text_style = {
'color': 'blue',
'va': 'bottom',
'ha': 'left',
'path_effects': [PathEffects.withStroke(linewidth=3, foreground="w")],
'fontsize': 14
}
text_ents = []
for i, player in enumerate(data.Name.unique()):
text_style['color'] = colour_list[i]
p_data = data.loc[data['Name'] == player]
text_ents += plot_labels(p_data, ax_in, text_style)
# text_ents = plot_labels(data, ax_in, text_style)
ward = Image_open(environment['WARD_ICON'])
ward.thumbnail(size)
extra_ents = plot_image_scatter(data, ax_in, ward)
# Adjust the labels
adjust_text(text_ents, extra_ents=extra_ents, ax=ax_in)
# Add summary table
data['label'] = [str(x + 1) for x in range(data.shape[0])]
table = plot_sum_table(data, ax_in)
extra_ents += text_ents
extra_ents.append(table)
return extra_ents
def loadImage(self, imageName):
glEnable(GL_TEXTURE_2D)
"""Load an image file as a 2D texture using PIL"""
im = ImageOpen(imageName)
img_data = numpy.array(im.getdata(), numpy.uint8)
self.imageID = glGenTextures(1)
print(self.imageID)
glPixelStorei(GL_UNPACK_ALIGNMENT, 1)
glBindTexture(GL_TEXTURE_2D, self.imageID)
glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP)
glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP)
glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_LINEAR)
glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR)
glTexImage2D(GL_TEXTURE_2D, 0, GL_RGB, im.size[0], im.size[1], 0,
GL_RGB, GL_UNSIGNED_BYTE, img_data)
return self.imageID
def loadImage( self, imageName):
glEnable(self.__texMode)
"""Load an image file as a 2D texture using PIL"""
im = ImageOpen(imageName)
img_data = numpy.array(im.getdata(), numpy.uint8)
self.imageID = glGenTextures(1)
print(self.imageID)
glPixelStorei(GL_UNPACK_ALIGNMENT,1)
glBindTexture(self.__texMode, self.imageID)
glTexParameterf(self.__texMode, GL_TEXTURE_WRAP_S, GL_REPEAT)
glTexParameterf(self.__texMode, GL_TEXTURE_WRAP_T, GL_REPEAT)
#glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_LINEAR)
#glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR)
glTexImage2D(
self.__texMode, 0, GL_RGB, im.size[0], im.size[1], 0,
GL_RGB, GL_UNSIGNED_BYTE, img_data
)
return self.imageID
def show_merge(self, flag='GT'):
if flag == 'GT':
path = self.path_mask_GT
plotwindow = self.merge
elif flag == 'pred':
path = self.path_mask_pred
plotwindow = self.merge_pred
(_, name) = os.path.split(self.image_list[self.current_image_mark])
try:
mask = Iopen(os.path.join(path, name))
image = Iopen(self.image_list[self.current_image_mark])
mask = mask.resize((511, 431))
image = image.resize((511, 431))
merge = array(image)
mask_arr = array(mask)
merge[mask_arr == 0] = [0, 0, 0]
merge = fromarray(merge)
plotwindow.setPixmap(merge.toqpixmap())
except FileNotFoundError as e:
pass
def load(path):
with open(path, 'rb') as f:
image = PIOpen(f)
image.load()
return VideoFrame(image)
def load(path):
with open(path, 'rb') as f:
image = PIOpen(f)
image.load()
return VideoFrame(image)
def __init__(self, path):
image = PIL_Image_open(path).convert("RGBA")
self.__size = math.Vector(image.size)
self.__data = image.tobytes("raw")
def __init__(self, addr):
# if addr!='':
self.__im = ImOpen(addr).convert('L')
class FVRT:
# 初始化各系数
# 变换后的宽
WIDTH = 75
# 变换后的高
HEIGHT = 170
# 灰度放大系数
THREFACTOR = 4
# 隶属度分割界限
THREGRAY = 0.3
# 迭代次数
ITER = 3
# Niblack模板边长
SIDELENGTH = 11
# 补偿权值
THREK = 0.05
# 开运算模板边长
OPENSIDE = 3
# 闭运算模板边长
CLOSEDSIDE = 3
# 中值滤波模板边长
BLURSIDE = 5
# 圈定面积
MINSIZE = 20
# 输入原始图像
def __init__(self, addr):
# if addr!='':
self.__im = ImOpen(addr).convert('L')
# else:
# self.__im=image
# 尺寸归一化
def SizeNorm(self, width, height):
self.__im = self.__im.resize((width, height), BILINEAR)
# 灰度归一化
def GrayNorm(self):
pix = array(self.__im)
norm = Normalize(vmin=pix.min(), vmax=pix.max())
pix = norm(pix) * 255
pix = pix.astype(int32)
self.__im = fromarray(pix)
# 基于方向的谷型检测
def DirectValley(self, ThreFactor):
pix = array(self.__im)
row = array([0, 2, 4, 6, 8])
value = array([3, -1, -4, -1, 3])
A = {}
for i in range(4):
if i == 0:
col = row
elif i == 1:
col = array([2, 3, 4, 5, 6])
elif i == 2:
col = array([4, 4, 4, 4, 4])
else:
col = array([6, 5, 4, 3, 2])
A[i] = csc_matrix((value, (row, col)), shape=(9, 9)).toarray()
A[i + 4] = rot90(A[i])
pixdict = {}
for i in range(8):
pixdict[i] = convolve2d(pix, A[i], 'same')
m, n = pix.shape
for x in range(m):
for y in range(n):
valmax = -1000
for i in range(8):
if pixdict[i][x, y] >= valmax:
valmax = pixdict[i][x, y]
if valmax >= 0:
pix[x, y] = valmax
else:
pix[x, y] = 0
nozero_num = count_nonzero(pix)
nozero_avg = sum(pix) / nozero_num
for x in range(m):
for y in range(n):
if pix[x, y] >= (ThreFactor * nozero_avg):
pix[x, y] = ThreFactor * nozero_avg
self.__im = fromarray(pix)
# 图像模糊增强
def BlurEnhance(self, ThreGray, Iter):
pix = array(self.__im)
pix = sin(pix * pi / (2 * 255))
for i in arange(Iter):
pix = (sin(pi * (pix - ThreGray)) + 1) / 2
pix = 255 * arcsin(pix**sqrt(Iter)) * 2 / pi
self.__im = fromarray(pix)
# 图像阈值分割
def ThreSegment(self, SideLength, ThreK):
pix = array(self.__im)
thresh_pix = threshold_niblack(pix,
window_size=SideLength * SideLength,
k=ThreK)
binary_pix = pix > thresh_pix
pix = binary_pix.astype(int) * 255
self.__im = fromarray(pix)
# 图像去噪
def Denoise(self, OpenSide, ClosedSide, BlurSide):
im = self.__im
im = cv2.cvtColor(asarray(im).astype(uint8), cv2.COLOR_GRAY2BGR)
openkernel = cv2.getStructuringElement(cv2.MORPH_RECT,
(OpenSide, OpenSide))
opening = cv2.morphologyEx(array(im), cv2.MORPH_OPEN, openkernel)
closedkernel = cv2.getStructuringElement(cv2.MORPH_RECT,
(ClosedSide, ClosedSide))
closed = cv2.morphologyEx(opening, cv2.MORPH_CLOSE, closedkernel)
result = cv2.medianBlur(closed, BlurSide)
pix = cv2.cvtColor(result, cv2.COLOR_BGR2GRAY)
self.__im = fromarray(pix)
# 图像细化
def Thin(self):
pix = array(self.__im)
pix = pix / 255
pix = morphology.skeletonize(pix)
pix = pix.astype(int) * 255
self.__im = fromarray(pix)
# 二值化图像去除残余斑点
def CleanSmallObject(self, MinSize):
pix = array(self.__im) / 255
pix = pix.astype(bool)
pix = morphology.remove_small_objects(pix,
min_size=MinSize,
connectivity=2)
pix = pix.astype(int) * 255
self.__im = fromarray(pix)
# 获得图像数组
def GetArray(self):
return array(self.__im)
# 获得图像
def GetImage(self):
return self.__im
# 显示图像
def ShowImage(self):
self.__im.show()
# 设置各系数
@staticmethod
def SetFactor(width=75,
height=170,
ThreFactor=4,
ThreGray=0.3,
Iter=3,
SideLength=11,
ThreK=0.05,
OpenSide=3,
ClosedSide=3,
BlurSide=5,
MinSize=20):
FVRT.WIDTH = width
FVRT.HEIGHT = height
FVRT.THREFACTOR = ThreFactor
FVRT.THREGRAY = ThreGray
FVRT.ITER = Iter
FVRT.SIDELENGTH = SideLength
FVRT.THREK = ThreK
FVRT.OPENSIDE = OpenSide
FVRT.CLOSEDSIDE = ClosedSide
FVRT.BLURSIDE = BlurSide
FVRT.MINSIZE = MinSize
# 一键操作图像
@staticmethod
def OneKey(addr):
image = FVRT(addr)
image.SizeNorm(FVRT.WIDTH, FVRT.HEIGHT)
image.GrayNorm()
image.DirectValley(FVRT.THREFACTOR)
image.BlurEnhance(FVRT.THREGRAY, FVRT.ITER)
image.ThreSegment(FVRT.SIDELENGTH, FVRT.THREK)
image.Denoise(FVRT.OPENSIDE, FVRT.CLOSEDSIDE, FVRT.BLURSIDE)
image.Thin()
image.CleanSmallObject(FVRT.MINSIZE)
return image.GetImage()
