def detect_image(self, image):
if self.model_image_size != (None, None):
assert self.model_image_size[0] % 32 == 0, 'Multiples of 32 required'
assert self.model_image_size[1] % 32 == 0, 'Multiples of 32 required'
boxed_image = image_preporcess(np.copy(image), tuple(reversed(self.model_image_size)))
image_data = boxed_image
out_boxes, out_scores, out_classes = self.sess.run(
[self.boxes, self.scores, self.classes],
feed_dict={
self.yolo_model.input: image_data,
self.input_image_shape: [image.shape[0], image.shape[1]], # [image.size[1], image.size[0]],
K.learning_phase(): 0
})
# print('Found {} boxes for {}'.format(len(out_boxes), 'img'))
thickness = (image.shape[0] + image.shape[1]) // 600
fontScale = 1
ObjectsList = []
for i, c in reversed(list(enumerate(out_classes))):
predicted_class = self.class_names[c]
box = out_boxes[i]
score = out_scores[i]
label = '{} {:.2f}'.format(predicted_class, score)
# label = '{}'.format(predicted_class)
scores = '{:.2f}'.format(score)
top, left, bottom, right = box
top = max(0, np.floor(top + 0.5).astype('int32'))
left = max(0, np.floor(left + 0.5).astype('int32'))
bottom = min(image.shape[0], np.floor(bottom + 0.5).astype('int32'))
right = min(image.shape[1], np.floor(right + 0.5).astype('int32'))
mid_h = (bottom - top) / 2 + top
mid_v = (right - left) / 2 + left
# put object rectangle
cv2.rectangle(image, (left, top), (right, bottom), self.colors[c], thickness)
# get text size
(test_width, text_height), baseline = cv2.getTextSize(label, cv2.FONT_HERSHEY_SIMPLEX,
thickness / self.text_size, 1)
# put text rectangle
cv2.rectangle(image, (left, top), (left + test_width, top - text_height - baseline), self.colors[c],
thickness=cv2.FILLED)
# put text above rectangle
cv2.putText(image, label, (left, top - 2), cv2.FONT_HERSHEY_SIMPLEX, thickness / self.text_size, (0, 0, 0),
1)
# add everything to list
ObjectsList.append([top, left, bottom, right, mid_v, mid_h, label, scores])
return image, ObjectsList
def detect_image_without_drawing(self, image):
if self.model_image_size != (None, None):
assert self.model_image_size[
0] % 32 == 0, 'Multiples of 32 required'
assert self.model_image_size[
1] % 32 == 0, 'Multiples of 32 required'
boxed_image = image_preporcess(
np.copy(image), tuple(reversed(self.model_image_size)))
image_data = boxed_image
out_boxes, out_scores, out_classes = self.sess.run(
[self.boxes, self.scores, self.classes],
feed_dict={
self.yolo_model.input: image_data,
self.input_image_shape:
[image.shape[0],
image.shape[1]], # [image.size[1], image.size[0]],
K.learning_phase(): 0
})
# print('Found {} boxes for {}'.format(len(out_boxes), 'img'))
thickness = (image.shape[0] + image.shape[1]) // 600
fontScale = 1
ObjectsList = []
for i, c in reversed(list(enumerate(out_classes))):
predicted_class = self.class_names[c]
box = out_boxes[i]
score = out_scores[i]
# label = '{} {:.2f}'.format(predicted_class, score)
# label = '{}'.format(predicted_class)
scores = '{:.2f}'.format(score)
top, left, bottom, right = box
top = max(0, np.floor(top + 0.5).astype('int32'))
left = max(0, np.floor(left + 0.5).astype('int32'))
bottom = min(image.shape[0],
np.floor(bottom + 0.5).astype('int32'))
right = min(image.shape[1], np.floor(right + 0.5).astype('int32'))
mid_h = (bottom - top) / 2 + top
mid_v = (right - left) / 2 + left
# add everything to list
ObjectsList.append(
[left, top, right, bottom, mid_v, mid_h, c, scores])
return ObjectsList
def detect_image(self, image):
# if self.model_image_size != (None, None):
# assert self.model_image_size[0] % 32 == 0, 'Multiples of 32 required'
# assert self.model_image_size[1] % 32 == 0, 'Multiples of 32 required'
# boxed_image = image_preporcess(np.copy(image), tuple(reversed(self.model_image_size)))
# image_data = boxed_image
# image_data = image/255
# image_data = np.expand_dims(image_data,0)
frame = cv2.resize(image, (416, 416))
if self.model_image_size != (None, None):
assert self.model_image_size[
0] % 32 == 0, 'Multiples of 32 required'
assert self.model_image_size[
1] % 32 == 0, 'Multiples of 32 required'
boxed_image = image_preporcess(
np.copy(frame), tuple(reversed(self.model_image_size)))
image_data = boxed_image
out_boxes, out_scores, out_classes = self.sess.run(
[self.boxes, self.scores, self.classes],
feed_dict={
self.yolo_model.input: image_data,
self.input_image_shape:
[frame.shape[0],
frame.shape[1]], # [image.size[1], image.size[0]],
K.learning_phase(): 0
})
# print('Found {} boxes for {}'.format(len(out_boxes), 'img'))
Rx = image.shape[1] / 416
Ry = image.shape[0] / 416
thickness = (image.shape[0] + image.shape[1]) // 600
fontScale = 1
ObjectsList = []
# print(out_classes)
for i, c in reversed(list(enumerate(out_classes))):
if c == 15:
continue
predicted_class = self.class_names[c]
box = out_boxes[i]
score = out_scores[i]
label = '{} {:.2f}'.format(predicted_class, score)
# label = '{}'.format(predicted_class)
scores = '{:.2f}'.format(score)
top, left, bottom, right = box
top = max(0, np.floor(top + 0.5).astype('int32'))
left = max(0, np.floor(left + 0.5).astype('int32'))
bottom = min(image.shape[0],
np.floor(bottom + 0.5).astype('int32'))
right = min(image.shape[1], np.floor(right + 0.5).astype('int32'))
top = int(top * Ry)
left = int(left * Rx)
bottom = int(bottom * Ry)
right = int(right * Rx)
if c == 4:
try:
img = image[top:bottom + 15, left - 20:right + 20]
text = read_plate.detect(img)
# text = Image_to_string(img)
text = text[:2] + '000'
cv2.putText(image, text[:5], (left, top),
cv2.FONT_HERSHEY_SIMPLEX, 2, (255, 255, 255),
3)
print(text)
except:
print('sth wrong')
# x, y, w, h = box
# print(box)
# top = max(0, np.floor(x + 0.5).astype(int))
# left = max(0, np.floor(y + 0.5).astype(int))
# right = min(image.shape[1], np.floor(x + w + 0.5).astype(int))
# bottom = min(image.shape[0], np.floor(y + h + 0.5).astype(int))
mid_h = (bottom - top) / 2 + top
mid_v = (right - left) / 2 + left
# put object rectangle
if c != 4:
cv2.rectangle(image, (left, top), (right, bottom),
self.colors[c], thickness)
# get text size
(test_width, text_height), baseline = cv2.getTextSize(
label, cv2.FONT_HERSHEY_SIMPLEX,
thickness / self.text_size, 1)
# put text rectangle
cv2.rectangle(
image, (left, top),
(left + test_width, top - text_height - baseline),
self.colors[c],
thickness=cv2.FILLED)
# put text above rectangle
cv2.putText(image, label, (left, top - 2),
cv2.FONT_HERSHEY_SIMPLEX,
thickness / self.text_size, (0, 0, 0), 1)
# add everything to list
ObjectsList.append(
[top, left, bottom, right, mid_v, mid_h, label, scores])
return image, ObjectsList
def detect_image(self, image):
# if self.model_image_size != (None, None):
# assert self.model_image_size[0] % 32 == 0, 'Multiples of 32 required'
# assert self.model_image_size[1] % 32 == 0, 'Multiples of 32 required'
# boxed_image = image_preporcess(np.copy(image), tuple(reversed(self.model_image_size)))
# image_data = boxed_image
# image_data = image/255
# image_data = np.expand_dims(image_data,0)
if self.model_image_size != (None, None):
assert self.model_image_size[
0] % 32 == 0, 'Multiples of 32 required'
assert self.model_image_size[
1] % 32 == 0, 'Multiples of 32 required'
boxed_image = image_preporcess(
np.copy(image), tuple(reversed(self.model_image_size)))
image_data = boxed_image
out_boxes, out_scores, out_classes = self.sess.run(
[self.boxes, self.scores, self.classes],
feed_dict={
self.yolo_model.input: image_data,
self.input_image_shape:
[image.shape[0],
image.shape[1]], # [image.size[1], image.size[0]],
K.learning_phase(): 0
})
# print('Found {} boxes for {}'.format(len(out_boxes), 'img'))
thickness = (image.shape[0] + image.shape[1]) // 600
fontScale = 1
ObjectsList = []
# print(out_classes)
left_axes = np.array([
max(0,
np.floor(out_boxes[i][1] + 0.5).astype('int32'))
for i, c in enumerate(out_classes) if c == 4
])
num_obj_true = 3 * len(left_axes)
if len(out_classes) < num_obj_true:
detect = 0
else:
detect = 1
no_obj = [str(i + self.num_obj) for i in range(len(left_axes))]
list_text = []
list_obj = []
pair = []
barcode_list = []
for i, c in enumerate(out_classes):
predicted_class = self.class_names[c]
box = out_boxes[i]
score = out_scores[i]
# label = '{} {:.2f}'.format(predicted_class, score)
label = '{}'.format(predicted_class)
scores = '{:.2f}'.format(score)
top, left, bottom, right = box
top = max(0, np.floor(top + 0.5).astype('int32'))
left = max(0, np.floor(left + 0.5).astype('int32'))
bottom = min(image.shape[0],
np.floor(bottom + 0.5).astype('int32'))
right = min(image.shape[1],
np.floor(right + 0.5).astype('int32'))
mid_h = (bottom - top) / 2 + top
mid_v = (right - left) / 2 + left
if left < 10 and (time.time() -
self.start_tracking) > 2 and i == 0:
self.num_obj += 1
self.start_tracking = time.time()
# print(self.num_obj)
if (c != 15) and (c != 4):
cv2.rectangle(image, (left, top), (right, bottom),
self.colors[c], thickness)
# (test_width, text_height), baseline = cv2.getTextSize(label, cv2.FONT_HERSHEY_SIMPLEX,
# thickness / self.text_size, 1)
#
# # put text rectangle
# cv2.rectangle(image, (left, top), (left + test_width, top - text_height - baseline), self.colors[c],
# thickness=cv2.FILLED)
#
# # put text above rectangle
# cv2.putText(image, label, (left, top - 2), cv2.FONT_HERSHEY_SIMPLEX, thickness / self.text_size,
# (0, 0, 0),1)
# distance = np.absolute(left - left_axes)
# print(distance)
distance = np.absolute(left - left_axes)
distance_index = np.where(distance < 200)
# print(distance_index)
distance_index = list(distance_index[0])
try:
del no_obj[distance_index[0]]
# print(left)
# print(list_text)
# print(list_text[distance_index[0]])
pair.append([
distance_index[0], label,
list_text[distance_index[0]]
])
except:
print('Something Wrong')
elif c == 4:
# having_text = 1
# image = scipy.ndimage.zoom(image,(2,2,1),order = 1)## upsampling image
try:
img = image[top:bottom + 15, left - 20:right + 20]
text = read_plate.detect(img)
n = len(text) - 3
text = text[:n] + '000'
list_text.append(text[:6])
cv2.putText(image, text[:6], (left, top),
cv2.FONT_HERSHEY_SIMPLEX, 2,
(255, 255, 255), 3)
except:
print('sth wrong in text')
elif c == 15:
try:
img = image[top:bottom + 15, left - 20:right + 20]
text = barcode_detect(img).decode("utf-8")
# cv2.putText(image, text[:6], (left, top), cv2.FONT_HERSHEY_SIMPLEX, 2, (255, 255, 255), 3)
# print(text)
barcode_list.append(text)
except:
print('sth wrong')
barcode_list.append('error')
#### hien thi text gia tren object
# print(text[:6])
not_match_list = []
# print(pair)
obj_text = make_no_obj_text(no_obj)
barcode_string = make_no_obj_text(barcode_list)
if len(no_obj) == 0:
obj_text = '...'
for idx, object, text in pair:
if check([object, text]):
continue
else:
not_match_list.append(str(idx + self.num_obj))
if len(not_match_list) == 0:
match_text = '...'
else:
match_text = make_no_obj_text(not_match_list)
# add everything to list
# ObjectsList.append([top, left, bottom, right, mid_v, mid_h, label, scores])
return image, detect, obj_text, match_text, no_obj, not_match_list, barcode_string
def detect_image(self, image):
if self.model_image_size != (None, None):
assert self.model_image_size[
0] % 32 == 0, 'Multiples of 32 required'
assert self.model_image_size[
1] % 32 == 0, 'Multiples of 32 required'
boxed_image = image_preporcess(
np.copy(image), tuple(reversed(self.model_image_size)))
image_data = boxed_image
out_boxes, out_scores, out_classes = self.sess.run(
[self.boxes, self.scores, self.classes],
feed_dict={
self.yolo_model.input: image_data,
self.input_image_shape:
[image.shape[0],
image.shape[1]], #[image.size[1], image.size[0]],
K.learning_phase(): 0
})
global second_pass
if second_pass == True:
yolo = YOLO()
thickness = (image.shape[0] + image.shape[1]) // 600
fontScale = 1
ObjectsList = []
global Motor
global Noframes
global frame
for i, c in reversed(list(enumerate(out_classes))):
predicted_class = self.class_names[c]
box = out_boxes[i]
score = out_scores[i]
#print((score))
print(predicted_class)
if (predicted_class == 'Hammer') or (predicted_class == 'Axe') or (
predicted_class
== 'Helmet') or (predicted_class == 'Kitchen_Knife') or (
predicted_class == 'Knife') or (predicted_class
== 'Handgun'):
print('&&&&&&&&&&&')
label = '{} {:.2f}'.format(predicted_class, score)
#label = '{}'.format(predicted_class)
scores = '{:.2f}'.format(score)
top, left, bottom, right = box
top = max(0, np.floor(top + 0.5).astype('int32'))
left = max(0, np.floor(left + 0.5).astype('int32'))
bottom = min(image.shape[0],
np.floor(bottom + 0.5).astype('int32'))
right = min(image.shape[1], np.floor(right + 0.5).astype('int32'))
mid_h = (bottom - top) / 2 + top
mid_v = (right - left) / 2 + left
global Helmet_Detected
if predicted_class == 'Axe':
print('abnormal')
elif predicted_class == 'Hammer':
print('abnormal')
elif predicted_class == 'Helmet':
print('abnormal')
elif predicted_class == 'Kitchen_Knife':
print('abnormal')
elif predicted_class == 'Knife':
print('abnormal')
return image, ObjectsList
def detect_image(self, image):
if self.model_image_size != (None, None):
assert self.model_image_size[
0] % 32 == 0, 'Multiples of 32 required'
assert self.model_image_size[
1] % 32 == 0, 'Multiples of 32 required'
boxed_image = image_preporcess(
np.copy(image), tuple(reversed(self.model_image_size)))
image_data = boxed_image
out_boxes, out_scores, out_classes = self.sess.run(
[self.boxes, self.scores, self.classes],
feed_dict={
self.yolo_model.input: image_data,
self.input_image_shape:
[image.shape[0],
image.shape[1]], #[image.size[1], image.size[0]],
K.learning_phase(): 0
})
#print('Found {} boxes for {}'.format(len(out_boxes), 'img'))
#thickness = (image.shape[0] + image.shape[1]) // 600
#fontScale=1
cropImgs = []
areas = []
for i, c in reversed(list(enumerate(out_classes))):
predicted_class = self.class_names[c]
box = out_boxes[i]
score = out_scores[i]
label = '{} {:.2f}'.format(predicted_class, score)
#label = '{}'.format(predicted_class)
#scores = '{:.2f}'.format(score)
top, left, bottom, right = box
top = max(0, np.floor(top + 0.5).astype('int32'))
left = max(0, np.floor(left + 0.5).astype('int32'))
bottom = min(image.shape[0],
np.floor(bottom + 0.5).astype('int32'))
right = min(image.shape[1], np.floor(right + 0.5).astype('int32'))
#mid_h = (bottom-top)/2+top
#mid_v = (right-left)/2+left
if str(label[0]) == "B":
# put object rectangle
if int(left - 20) < 0:
boxLeft = left
else:
boxLeft = int(left - 20)
if int(top - 20) < 0:
boxTop = top
else:
boxTop = int(top - 20)
if int(right + 20) > image.shape[1]:
boxRight = right
else:
boxRight = int(right + 20)
if int(bottom + 20) > image.shape[0]:
boxBottom = bottom
else:
boxBottom = int(bottom + 20)
#cv2.rectangle(image, (boxLeft, boxTop), (boxRight, boxBottom), self.colors[c], thickness)
#img = cv2.imread(image)
crop_img = image[boxTop:boxBottom, boxLeft:boxRight]
#area = (bottom - top)*(right - left)
area = (boxBottom - boxTop) * (boxRight - boxLeft)
areas.append(area)
cropImgs.append(crop_img)
else:
pass
# get text size
#(test_width, text_height), baseline = cv2.getTextSize(label, cv2.FONT_HERSHEY_SIMPLEX, thickness/self.text_size, 1)
# put text rectangle
#cv2.rectangle(image, (left, top), (left + test_width, top - text_height - baseline), self.colors[c], thickness=cv2.FILLED)
# put text above rectangle
#cv2.putText(image, label, (left, top-2), cv2.FONT_HERSHEY_SIMPLEX, thickness/self.text_size, (0, 0, 0), 1)
# add everything to list
try:
crop = cropImgs[areas.index(max(areas))]
except:
crop = -1
return image, crop
def detect_image(self, image):
# if self.model_image_size != (None, None):
# assert self.model_image_size[0] % 32 == 0, 'Multiples of 32 required'
# assert self.model_image_size[1] % 32 == 0, 'Multiples of 32 required'
# boxed_image = image_preporcess(np.copy(image), tuple(reversed(self.model_image_size)))
# image_data = boxed_image
# image_data = image/255
# image_data = np.expand_dims(image_data,0)
frame = cv2.resize(image, (416, 416))
if self.model_image_size != (None, None):
assert self.model_image_size[
0] % 32 == 0, 'Multiples of 32 required'
assert self.model_image_size[
1] % 32 == 0, 'Multiples of 32 required'
boxed_image = image_preporcess(
np.copy(frame), tuple(reversed(self.model_image_size)))
image_data = boxed_image
out_boxes, out_scores, out_classes = self.sess.run(
[self.boxes, self.scores, self.classes],
feed_dict={
self.yolo_model.input: image_data,
self.input_image_shape:
[frame.shape[0],
frame.shape[1]], # [image.size[1], image.size[0]],
K.learning_phase(): 0
})
# print('Found {} boxes for {}'.format(len(out_boxes), 'img'))
Rx = image.shape[1] / 416
Ry = image.shape[0] / 416
left_text = top_text = width_text = height_text = _conf_text = 0
_class_text = None
for i, c in reversed(list(enumerate(out_classes))):
predicted_class = self.class_names[c]
box = out_boxes[i]
score = out_scores[i]
label = '{} {:.2f}'.format(predicted_class, score)
# label = '{}'.format(predicted_class)
scores = '{:.2f}'.format(score)
top, left, bottom, right = box
top = max(0, np.floor(top + 0.5).astype('int32'))
left = max(0, np.floor(left + 0.5).astype('int32'))
bottom = min(image.shape[0],
np.floor(bottom + 0.5).astype('int32'))
right = min(image.shape[1], np.floor(right + 0.5).astype('int32'))
top = int(top * Ry)
left = int(left * Rx)
bottom = int(bottom * Ry)
right = int(right * Rx)
if c == 14:
# put object rectangle
left_text = left
top_text = top
right_text = right
bottom_text = bottom
_class_text = c
_conf_text = score
# image = frame[top:top + height, left:left + width] # crop image to recognize
# ve bounding boxes
# drawPred(classIDs[i], confidences[i], left, top, left + width, top + height)
# if len(indices) and having_text != 1:
# detect = 1
# having_text = 0
# else:
# detect = 0
print(_class_text)
# print(len(indices))
if _class_text == None:
return [_class_text]
else:
return [
_class_text, left_text, top_text, right_text, bottom_text,
_conf_text
]
def detect_image(self, image):
# if self.model_image_size != (None, None):
# assert self.model_image_size[0] % 32 == 0, 'Multiples of 32 required'
# assert self.model_image_size[1] % 32 == 0, 'Multiples of 32 required'
# boxed_image = image_preporcess(np.copy(image), tuple(reversed(self.model_image_size)))
# image_data = boxed_image
# image_data = image/255
# image_data = np.expand_dims(image_data,0)
frame = cv2.resize(image, (416, 416))
if self.model_image_size != (None, None):
assert self.model_image_size[
0] % 32 == 0, 'Multiples of 32 required'
assert self.model_image_size[
1] % 32 == 0, 'Multiples of 32 required'
boxed_image = image_preporcess(
np.copy(frame), tuple(reversed(self.model_image_size)))
image_data = boxed_image
out_boxes, out_scores, out_classes = self.sess.run(
[self.boxes, self.scores, self.classes],
feed_dict={
self.yolo_model.input: image_data,
self.input_image_shape:
[frame.shape[0],
frame.shape[1]], # [image.size[1], image.size[0]],
K.learning_phase(): 0
})
# print('Found {} boxes for {}'.format(len(out_boxes), 'img'))
Rx = image.shape[1] / 416
Ry = image.shape[0] / 416
thickness = (image.shape[0] + image.shape[1]) // 600
fontScale = 1
ObjectsList = []
_class_text = None
for i, c in reversed(list(enumerate(out_classes))):
predicted_class = self.class_names[c]
box = out_boxes[i]
score = out_scores[i]
label = '{} {:.2f}'.format(predicted_class, score)
# label = '{}'.format(predicted_class)
scores = '{:.2f}'.format(score)
top, left, bottom, right = box
top = max(0, np.floor(top + 0.5).astype('int32'))
left = max(0, np.floor(left + 0.5).astype('int32'))
bottom = min(image.shape[0],
np.floor(bottom + 0.5).astype('int32'))
right = min(image.shape[1], np.floor(right + 0.5).astype('int32'))
top = int(top * Ry)
left = int(left * Rx)
bottom = int(bottom * Ry)
right = int(right * Rx)
mid_h = (bottom - top) / 2 + top
mid_v = (right - left) / 2 + left
# put object rectangle
cv2.rectangle(image, (left, top), (right, bottom), self.colors[c],
thickness)
# get text size
(test_width, text_height), baseline = cv2.getTextSize(
label, cv2.FONT_HERSHEY_SIMPLEX, thickness / self.text_size, 1)
# put text rectangle
cv2.rectangle(image, (left, top),
(left + test_width, top - text_height - baseline),
self.colors[c],
thickness=cv2.FILLED)
# put text above rectangle
cv2.putText(image, label, (left, top - 2),
cv2.FONT_HERSHEY_SIMPLEX, thickness / self.text_size,
(0, 0, 0), 1)
# add everything to list
if c == 11:
left_text = left
top_text = top
bottom_text = bottom
right_text = right
_class_text = c
if _class_text == None:
return [_class_text]
else:
return [_class_text, left_text, top_text, right_text, bottom_text]