def binstatus2(items):
while items != [0,0,0,0]:
open = door()
#for each element of the items array
for i in range(1, 4):
#check if the object is equal to the bottom compartment category
if items[i-1] == i:
platenumber = i
#if object matches compartment category, empty the compartment
openplate()
items[i] = 0
rotate()
elif:
rotate()
if open == True:
break
return(items)
def next_batch(self, batch_size, phase):
# Get next batch of image (path) and labels
if phase == 'train':
if self.train_ptr + batch_size < self.train_size:
paths = self.train_image[self.train_ptr:self.train_ptr + batch_size]
labels = self.train_label[self.train_ptr:self.train_ptr + batch_size]
self.train_ptr += batch_size
else:
new_ptr = (self.train_ptr + batch_size)%self.train_size
paths = self.train_image[self.train_ptr:] + self.train_image[:new_ptr]
labels = self.train_label[self.train_ptr:] + self.train_label[:new_ptr]
self.train_ptr = new_ptr
elif phase == 'test':
if self.test_ptr + batch_size < self.test_size:
paths = self.test_image[self.test_ptr:self.test_ptr + batch_size]
labels = self.test_label[self.test_ptr:self.test_ptr + batch_size]
self.test_ptr += batch_size
else:
new_ptr = (self.test_ptr + batch_size)%self.test_size
paths = self.test_image[self.test_ptr:] + self.test_image[:new_ptr]
labels = self.test_label[self.test_ptr:] + self.test_label[:new_ptr]
self.test_ptr = new_ptr
else:
return None, None
# Read images
images = np.ndarray([batch_size, self.crop_size, self.crop_size, 3])
for i in xrange(len(paths)):
img = cv2.imread(paths[i])
h, w, c = img.shape
assert c==3
if phase == 'train':
angle = np.random.rand()*20 -10
img = rotate(img, angle)
img = cv2.resize(img, (self.scale_size, self.scale_size))
img = img.astype(np.float32)
img -= self.mean
# shift = int((self.scale_size-self.crop_size)/2)
shift1 = np.random.randint(0,self.scale_size-self.crop_size+1)
shift2 = np.random.randint(0,self.scale_size-self.crop_size+1)
img_crop = img[shift1:shift1+self.crop_size, shift2:shift2+self.crop_size, :]
images[i] = img_crop
else:
img = cv2.resize(img, (self.crop_size, self.crop_size))
img = img.astype(np.float32)
img -= self.mean
images[i] = img
# Expand labels
one_hot_labels = np.zeros((batch_size, self.n_classes))
for i in xrange(len(labels)):
one_hot_labels[i][labels[i]] = 1
return images, one_hot_labels
def STA(funfcn, Best, SE, Range, Iterations):
alpha_max = 1
alpha_min = 1e-4
alpha = alpha_max
beta = 1
gamma = 1
delta = 1
fc = 2
history = np.empty((Iterations, 1))
fBest = funfcn(Best[0]) # 用一种奇怪的方式调用矩阵中的数
for iter in range(Iterations):
Best, fBest = expand(funfcn, Best, fBest, SE, Range, beta, gamma)
Best, fBest = rotate(funfcn, Best, fBest, SE, Range, alpha, beta)
Best, fBest = axesion(funfcn, Best, fBest, SE, Range, beta, delta)
history[iter] = fBest
alpha = alpha / fc if alpha > alpha_min else alpha_max
return Best, fBest, history
_, head, _, tail = car
# park小车车
if park(head, destination.des, 70) == 'P':
comBlutooth('P', port)
comBlutooth('S', port)
break
# rotate小车车
th = thin(img)
cv2.imshow('thin', th)
op, l, m, r = rotate(th, 40,22, tail, head, oldOp)
oldOp = op
comBlutooth(op, port)
cv2.circle(img, tuple(head), 3, (0, 255, 0), -1)
cv2.circle(img, tuple(tail), 3, (255, 225, 0), -1)
cv2.circle(img, tuple(l), 5, (127, 0, 127), -1)
cv2.circle(img, tuple(m), 5, (0, 127, 127), -1)
cv2.circle(img, tuple(r), 5, (127, 127, 0), -1)
cv2.imshow("l & m & r & head & tail", img)
if cv2.waitKey(1) == ord('q'):
comBlutooth('P', port)
comBlutooth('S', port)
def teach(problem, solution, step):
step_now = solution.split(' ')[step]
config = [problem[i * 9:(i + 1) * 9] for i in range(6)]
config = rotate(config, step_now)
problem = ('').join(config)
return problem
# ############## img的canny变换 @ input:img output:edges img并没有变化
# cv2.imshow('edges', edges)
# ############## img的canny变换 @ input:img output:edges img并没有变化
undistorted_img = img[0:90, 0:90]
cv2.imshow('undistorted_img', undistorted_img)
if img[45 - kuandu:45 + kuandu,
0:90].sum() / 255 - 2 * kuandu * 90 * baifenbi >= 0 and img[
0:90, 45 - kuandu:45 +
kuandu].sum() / 255 - 2 * kuandu * 90 * baifenbi:
shiziflag = 1
else:
shiziflag = 0
# *** *** *** *** 十字检测任何时刻都在检测，没有flag
# ————————————————————————————————————————————————————————————————————
loc[1], shuzhe[1] = guoxian(undistorted_img, 0, shuzhe[1])
loc[2], shuzhe[2] = guoxian(rotate(undistorted_img, 10), 0, shuzhe[2])
loc[3], shuzhe[3] = guoxian(rotate(undistorted_img, -10), 0, shuzhe[3])
shuzhe[0][0] = shuzhe[0][1]
shuzhe[0][1] = shuzhe[0][2]
shuzhe[0][2] = max(shuzhe[1][2], shuzhe[2][2], shuzhe[3][2])
finalloc = []
loc[0] = [loc[1], loc[2], loc[3]]
for i in loc[0]:
if str(i) == 'nan':
continue
else:
finalloc.append(i)
if len(finalloc) == 0:
finalloc = np.nan
lower = np.array(lower, dtype="uint8")
upper = np.array(upper, dtype="uint8")
mask = cv2.inRange(img, lower, upper)
img = cv2.bitwise_and(img, img, mask=mask)
# ############## img的色彩提取 @ input:img output:img
img = cv2.resize(img, DIM)
map1, map2 = cv2.fisheye.initUndistortRectifyMap(K, D, np.eye(3), K, DIM,
cv2.CV_16SC2)
img = cv2.remap(img,
map1,
map2,
interpolation=cv2.INTER_LINEAR,
borderMode=cv2.BORDER_CONSTANT)
# ############## img的畸变矫正 @ input:img output:img
cv2.imshow('img', img)
cv2.imshow('rotate', rotate(img, 10))
# cv2.imshow('imgcopy', img[0:160, 0:90])
# img = img[0:90, 0:90]
img = rotate(img, 10)
img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
wtf, img = cv2.threshold(img, 120, 255, cv2.THRESH_BINARY)
# cv2.imshow('erzhi', img)
# ############## img灰度化加二值化 # input:img output:img
cany = cv2.GaussianBlur(img, (13, 13), 0)
edges = cv2.Canny(cany, 50, 70, apertureSize=3)
# ############## img的canny变换 @ input:img output:edges img并没有变化
cv2.imshow('edges', edges)
# ############## img的canny变换 @ input:img output:edges img并没有变化
undistorted_img = img
cv2.imshow('shizi', img)
if img[45 - kuandu:45 + kuandu,