def read_flow(flow_or_path, quantize=False, *args, **kwargs):
"""Read an optical flow map
Args:
flow_or_path(ndarray or str): either a flow map or path of a flow
quantize(bool): whether to read quantized pair, if set to True,
remaining args will be passed to :func:`dequantize_flow`
Returns:
ndarray: optical flow
"""
if isinstance(flow_or_path, np.ndarray):
if (flow_or_path.ndim != 3) or (flow_or_path.shape[-1] != 2):
raise ValueError(
'Invalid flow with shape {}'.format(flow_or_path.shape))
return flow_or_path
elif not isinstance(flow_or_path, str):
raise TypeError(
'"flow_or_path" must be a filename or numpy array, not {}'.format(
type(flow_or_path)))
if not quantize:
with open(flow_or_path, 'rb') as f:
try:
header = f.read(4).decode('utf-8')
except:
raise IOError('Invalid flow file: {}'.format(flow_or_path))
else:
if header != 'PIEH':
raise IOError(
'Invalid flow file: {}, header does not contain PIEH'.
format(flow_or_path))
w = np.fromfile(f, np.int32, 1).squeeze()
h = np.fromfile(f, np.int32, 1).squeeze()
flow = np.fromfile(f, np.float32, w * h * 2).reshape((h, w, 2))
else:
from cvbase.image import read_img
from cvbase.opencv import IMREAD_UNCHANGED
dx_filename, dy_filename = _pair_name(flow_or_path)
dx = read_img(dx_filename, flag=IMREAD_UNCHANGED)
dy = read_img(dy_filename, flag=IMREAD_UNCHANGED)
flow = dequantize_flow(dx, dy, *args, **kwargs)
return flow.astype(np.float32)
def show_img(img, win_name='', wait_time=0):
"""Show an image
Args:
img(str or ndarray): the image to be shown
win_name(str): the window name
wait_time(int): value of waitKey param
"""
cv2.imshow(win_name, read_img(img))
cv2.waitKey(wait_time)
def draw_bboxes_with_label(img, bboxes, labels, top_k=0, bbox_color=Color.green,
text_color=Color.green, thickness=1, font_scale=0.5,
show=True, win_name='', wait_time=0, out_file=None): # yapf: disable
"""Draw bboxes with label text in image
Args:
img(str or ndarray): the image to be shown
bboxes(list or ndarray): a list of ndarray of shape (k, 4)
labels(str or list): label name file or list of label names
top_k(int): draw top_k bboxes only if positive
bbox_color(Color or tuple): color to draw bboxes
text_color(Color or tuple): color to draw label texts
thickness(int): thickness of bbox lines
font_scale(float): font scales
show(bool): whether to show the image
win_name(str): the window name
wait_time(int): value of waitKey param
out_file(str or None): the filename to write the image
"""
img = read_img(img)
label_names = read_labels(labels)
if isinstance(bbox_color, Color):
bbox_color = bbox_color.value
if isinstance(text_color, Color):
text_color = text_color.value
assert len(bboxes) == len(label_names)
for i, _bboxes in enumerate(bboxes):
bboxes_int = _bboxes[:, :4].astype(np.int32)
if top_k <= 0:
_top_k = _bboxes.shape[0]
else:
_top_k = min(top_k, _bboxes.shape[0])
for j in range(_top_k):
left_top = (bboxes_int[j, 0], bboxes_int[j, 1])
right_bottom = (bboxes_int[j, 2], bboxes_int[j, 3])
cv2.rectangle(img,
left_top,
right_bottom,
bbox_color,
thickness=thickness)
if _bboxes.shape[1] > 4:
label_text = '{}|{:.02f}'.format(label_names[i], _bboxes[j, 4])
else:
label_text = label_names[i]
cv2.putText(img, label_text,
(bboxes_int[j, 0], bboxes_int[j, 1] - 2),
cv2.FONT_HERSHEY_COMPLEX, font_scale, text_color)
if show:
show_img(img, win_name, wait_time)
if out_file is not None:
write_img(img, out_file)
def test_resize_keep_ar(self):
# resize (400, 300) to (max_1000, max_600)
resized_img = resize_keep_ar(self.img_path, 1000, 600)
assert resized_img.shape == (600, 800, 3)
resized_img, scale = resize_keep_ar(self.img_path, 1000, 600, True)
assert resized_img.shape == (600, 800, 3) and scale == 2.0
# resize (400, 300) to (max_200, max_180)
img = read_img(self.img_path)
resized_img = resize_keep_ar(img, 200, 180)
assert resized_img.shape == (150, 200, 3)
resized_img, scale = resize_keep_ar(self.img_path, 200, 180, True)
assert resized_img.shape == (150, 200, 3) and scale == 0.5
# max_long_edge cannot be less than max_short_edge
with pytest.raises(ValueError):
resize_keep_ar(self.img_path, 500, 600)
def test_limit_size(self):
# limit to 800
resized_img = limit_size(self.img_path, 800)
assert resized_img.shape == (300, 400, 3)
resized_img, scale = limit_size(self.img_path, 800, True)
assert resized_img.shape == (300, 400, 3) and scale == 1
# limit to 200
resized_img = limit_size(self.img_path, 200)
assert resized_img.shape == (150, 200, 3)
resized_img, scale = limit_size(self.img_path, 200, True)
assert resized_img.shape == (150, 200, 3) and scale == 0.5
# test with img rather than img path
img = read_img(self.img_path)
resized_img = limit_size(img, 200)
assert resized_img.shape == (150, 200, 3)
resized_img, scale = limit_size(img, 200, True)
assert resized_img.shape == (150, 200, 3) and scale == 0.5
def draw_bboxes(img, bboxes, colors=Color.green, top_k=0, thickness=1,
show=True, win_name='', wait_time=0, out_file=None): # yapf: disable
"""Draw bboxes on an image
Args:
img(str or ndarray): the image to be shown
bboxes(list or ndarray): a list of ndarray of shape (k, 4)
colors(list or Color or tuple): a list of colors, corresponding to bboxes
top_k(int): draw top_k bboxes only if positive
thickness(int): thickness of lines
show(bool): whether to show the image
win_name(str): the window name
wait_time(int): value of waitKey param
out_file(str or None): the filename to write the image
"""
img = read_img(img)
if isinstance(bboxes, np.ndarray):
bboxes = [bboxes]
if isinstance(colors, (tuple, Color)):
colors = [colors for _ in range(len(bboxes))]
for i in range(len(colors)):
if isinstance(colors[i], Color):
colors[i] = colors[i].value
assert len(bboxes) == len(colors)
for i, _bboxes in enumerate(bboxes):
_bboxes = _bboxes.astype(np.int32)
if top_k <= 0:
_top_k = _bboxes.shape[0]
else:
_top_k = min(top_k, _bboxes.shape[0])
for j in range(_top_k):
left_top = (_bboxes[j, 0], _bboxes[j, 1])
right_bottom = (_bboxes[j, 2], _bboxes[j, 3])
cv2.rectangle(img,
left_top,
right_bottom,
colors[i],
thickness=thickness)
if show:
show_img(img, win_name, wait_time)
if out_file is not None:
write_img(img, out_file)
def test_read_img(self):
img = read_img(self.img_path)
assert img.shape == (300, 400, 3)
