def test_mean_image_wrong_parameters_4():
try:
phm_img.mean_image([[]])
except Exception as e:
assert type(e) == TypeError
else:
assert False
def test_mean_image_wrong_parameters_2():
try:
phm_img.mean_image(list())
except Exception as e:
assert type(e) == ValueError
else:
assert False
def test_mean_image_wrong_parameters_6():
try:
image = numpy.zeros((25, 25, 3))
phm_img.mean_image(image)
except Exception as e:
assert type(e) == TypeError
else:
assert False
def test_mean_image_wrong_parameters_7():
images = list()
images.append(numpy.ones((25, 25, 3)))
images.append(numpy.zeros((15, 15, 3)))
try:
phm_img.mean_image(images)
except Exception, e:
assert type(e) == ValueError
def test_mean_image_1():
images = list()
for i in range(10):
images.append(numpy.zeros((25, 25, 3)))
image = phm_img.mean_image(images)
assert numpy.count_nonzero(image) == 0
assert image.ndim == 3
assert image.shape == (25, 25, 3)
def test_threshold_meanshift_4():
im1 = numpy.zeros((25, 25, 3), numpy.uint8)
im2 = numpy.zeros((25, 25, 3), numpy.uint8)
mask = numpy.zeros((25, 25), numpy.uint8)
mean_im = phm_img.mean_image(list([im1, im2]))
bin_img = phm_img.threshold_meanshift(im1, mean_im, mask=mask, reverse=True)
assert bin_img.shape == (25, 25)
def test_threshold_meanshift_1():
im1 = numpy.zeros((25, 25, 3), numpy.uint8)
im2 = numpy.zeros((25, 25, 3), numpy.uint8)
mean_im = phm_img.mean_image(list([im1, im2]))
bin_img = phm_img.threshold_meanshift(im1, mean_im)
assert bin_img.shape == (25, 25)
def test_mean_image_3():
images = list()
for i in range(0, 1):
images.append(numpy.ones((25, 25, 3)))
for i in range(1, 10):
images.append(numpy.zeros((25, 25, 3)))
image = phm_img.mean_image(images)
assert (image == 0.1).all()
assert image.ndim == 3
assert image.shape == (25, 25, 3)
def test_threshold_meanshift_no_regression_1():
plant_number = 6
raw_side_images = phm_data.raw_images(plant_number=plant_number)['side']
mask = phm_data.tutorial_data_binarization_mask()[1]
mean_im = phm_img.mean_image(raw_side_images.values())
im_bin = phm_img.threshold_meanshift(raw_side_images[0], mean_im, mask=mask)
ref = 158482
# Acceptation error of 0.01 %
acceptation_error = ref * 0.001
if abs(numpy.count_nonzero(im_bin) - ref) > acceptation_error:
assert False
def binarize(raw_images):
# Compute the mean image of the side view image
mean_img = phm_img.mean_image(raw_images['side'].values())
routine_binarization = {
'side': lambda im: routine_side_binarization(im, mean_img),
'top': lambda im: routine_top_binarization(im)
}
bin_images = collections.defaultdict(dict)
for id_camera in raw_images:
for angle in raw_images[id_camera]:
bin_images[id_camera][angle] = routine_binarization[id_camera](
raw_images[id_camera][angle])
return bin_images
def test_mean_image_wrong_parameters_1():
try:
phm_img.mean_image(None)
except Exception, e:
assert type(e) == TypeError
# Convert image on HSV representation
hsv_image = cv2.cvtColor(image, cv2.COLOR_BGR2HSV)
# Apply a median blur on the image
hsv_image = cv2.medianBlur(hsv_image, ksize=median_blur_size)
# Threshold the image with HSV min and max value
bin_img = phm_img.threshold_hsv(hsv_image, hsv_min, hsv_max)
# dilate and erode the image to remove possible noise
bin_img = phm_img.dilate_erode(bin_img, kernel_shape=(3, 3),iterations=iterations)
return bin_img
## binarize image
# Compute the mean image of the side view image
mean_img = phm_img.mean_image(raw_images['side'].values())
routine_binarization = {
'side': lambda im : routine_side_binarization(im, mean_img),
'top': lambda im : routine_top_binarization(im)}
bin_images = collections.defaultdict(dict)
for id_camera in raw_images:
for angle in raw_images[id_camera]:
bin_images[id_camera][angle] = routine_binarization[id_camera](raw_images[id_camera][angle])
## Display images binarize
id_camera, angle = 'side', 120
phm_display.show_images([raw_images[id_camera][angle], mean_img, bin_images[id_camera][angle]])
id_camera, angle = 'top', 0
