def psIImask(img, mode='thresh'):
# pcv.plot_image(img)
if mode is 'thresh':
# this entropy based technique seems to work well when algae is present
algaethresh = filters.threshold_yen(image=img)
threshy = pcv.threshold.binary(img, algaethresh, 255, 'light')
# mask = pcv.dilate(threshy, 2, 1)
mask = pcv.fill(threshy, 250)
mask = pcv.erode(mask, 2, 1)
mask = pcv.fill(mask, 100)
final_mask = mask # pcv.fill(mask, 270)
elif isinstance(mode, pd.DataFrame):
mode = curvedf
rownum = mode.imageid.values.argmax()
imgdf = mode.iloc[[1, rownum]]
fm = cv2.imread(imgdf.filename[0])
fmp = cv2.imread(imgdf.filename[1])
npq = np.float32(np.divide(fm, fmp, where=fmp != 0) - 1)
npq = np.ma.array(fmp, mask=fmp < 200)
plt.imshow(npq)
# pcv.plot_image(npq)
final_mask = np.zeros_like(img)
else:
pcv.fatal_error(
'mode must be "thresh" (default) or an object of class pd.DataFrame'
)
return final_mask
def psIImask(img, mode='thresh'):
'''
Input:
img = greyscale image
mode = type of thresholding to perform. Currently only 'thresh' is available
'''
# pcv.plot_image(img)
if mode is 'thresh':
# this entropy based technique seems to work well when algae is present
algaethresh = filters.threshold_yen(image=img)
threshy = pcv.threshold.binary(img, algaethresh, 255, 'light')
# mask = pcv.dilate(threshy, 2, 1)
mask = pcv.fill(threshy, 150)
mask = pcv.erode(mask, 2, 1)
mask = pcv.fill(mask, 45)
# mask = pcv.dilate(mask, 2,1)
final_mask = mask # pcv.fill(mask, 270)
else:
pcv.fatal_error(
'mode must be "thresh" (default) or an object of class pd.DataFrame'
)
return final_mask
def local_orientation_label_cost(labeled_lines,
labeled_lines_num,
intact_lines_num,
max_orientation,
max_response,
theta,
radius_constant=18):
pixel_list = regionprops(np.transpose(labeled_lines))
label_cost = np.zeros((labeled_lines_num + 1, 1))
local_max_orientation = np.zeros((labeled_lines_num, 1))
logical = labeled_lines > 0
logical_double = logical.astype(np.double)
# TODO decide number of iterations
border_mask = np.logical_and(logical,
np.logical_not(erode(logical_double, 3, 1)))
border_mask = border_mask.astype(np.double)
# area divide by perimeter
sw = np.sum(logical_double) / np.sum(border_mask)
se = int(round(sw * radius_constant))
line_theta = np.zeros((labeled_lines_num, 1))
for i in range(intact_lines_num, labeled_lines_num):
x = pixel_list[i].coords
try:
pca = PCA()
pca_res = pca.fit(x)
pcav = pca_res.components_[0]
line_theta[i] = math.atan(pcav[1] / pcav[0])
except Exception as e:
line_theta[i] = np.inf
continue
max_row_coord = np.amax(x, 0)[1] + se
min_row_coord = max(np.amin(x, 0)[1] - se, 0)
max_col_coord = np.amax(x, 0)[0] + se
min_col_coord = max(np.amin(x, 0)[0] - se, 0)
roi = labeled_lines[min_row_coord:max_row_coord,
min_col_coord:max_col_coord]
logical_roi = roi == pixel_list[i].label
# logical = labeled_lines == i + 1
logical_double = logical_roi.astype(np.double)
mask = np.full(labeled_lines.shape, False)
# TODO number of iterations
mask[min_row_coord:max_row_coord,
min_col_coord:max_col_coord] = dilate(logical_double, se, 1)
# mask = dilate(logical_double, se, 1)
res = estimate_local_orientations(max_orientation, max_response, theta,
mask)
index = np.argmax(res[:, 1])
local_max_orientation[i] = res[index, 0]
label_cost[i] = 10 * np.exp(50 * (1 - abs(
np.cos(math.radians(local_max_orientation[i]) - line_theta[i]))))
return label_cost
def main():
# Create input arguments object
args = options()
# Set debug mode
pcv.params.debug = args.debug
# Open a single image
img, imgpath, imgname = pcv.readimage(filename=args.image)
# Visualize colorspaces
all_cs = pcv.visualize.colorspaces(rgb_img=img)
# Extract the Blue-Yellow ("b") channel from the LAB colorspace
gray_img = pcv.rgb2gray_lab(rgb_img=img, channel="b")
# Plot a histogram of pixel values for the Blue-Yellow ("b") channel.
hist_plot = pcv.visualize.histogram(gray_img=gray_img)
# Apply a binary threshold to the Blue-Yellow ("b") grayscale image.
thresh_img = pcv.threshold.binary(gray_img=gray_img,
threshold=140,
max_value=255,
object_type="light")
# Apply a dilation with a 5x5 kernel and 3 iterations
dil_img = pcv.dilate(gray_img=thresh_img, ksize=5, i=3)
# Fill in small holes in the leaves
closed_img = pcv.fill_holes(bin_img=dil_img)
# Erode the plant pixels using a 5x5 kernel and 3 iterations
er_img = pcv.erode(gray_img=closed_img, ksize=5, i=3)
# Apply a Gaussian blur with a 5 x 5 kernel.
blur_img = pcv.gaussian_blur(img=er_img, ksize=(5, 5))
# Set pixel values less than 255 to 0
blur_img[np.where(blur_img < 255)] = 0
# Fill/remove objects less than 300 pixels in area
cleaned = pcv.fill(bin_img=blur_img, size=300)
# Create a circular ROI
roi, roi_str = pcv.roi.circle(img=img, x=1725, y=1155, r=400)
# Identify objects in the binary image
cnts, cnts_str = pcv.find_objects(img=img, mask=cleaned)
# Filter objects by region of interest
plant_cnt, plant_str, plant_mask, plant_area = pcv.roi_objects(
img=img,
roi_contour=roi,
roi_hierarchy=roi_str,
object_contour=cnts,
obj_hierarchy=cnts_str)
# Combine objects into one
plant, mask = pcv.object_composition(img=img,
contours=plant_cnt,
hierarchy=plant_str)
# Measure size and shape properties
shape_img = pcv.analyze_object(img=img, obj=plant, mask=mask)
if args.writeimg:
pcv.print_image(img=shape_img,
filename=os.path.join(args.outdir,
"shapes_" + imgname))
# Analyze color properties
color_img = pcv.analyze_color(rgb_img=img, mask=mask, hist_plot_type="hsv")
if args.writeimg:
pcv.print_image(img=color_img,
filename=os.path.join(args.outdir,
"histogram_" + imgname))
# Save the measurements to a file
pcv.print_results(filename=args.result)
def check_cycles(skel_img):
""" Check for cycles in a skeleton image
Inputs:
skel_img = Skeletonized image
Returns:
cycle_img = Image with cycles identified
:param skel_img: numpy.ndarray
:return cycle_img: numpy.ndarray
"""
# Store debug
debug = params.debug
params.debug = None
# Create the mask needed for cv2.floodFill, must be larger than the image
h, w = skel_img.shape[:2]
mask = np.zeros((h + 2, w + 2), np.uint8)
# Copy the skeleton since cv2.floodFill will draw on it
skel_copy = skel_img.copy()
cv2.floodFill(skel_copy, mask=mask, seedPoint=(0, 0), newVal=255)
# Invert so the holes are white and background black
just_cycles = cv2.bitwise_not(skel_copy)
# Erode slightly so that cv2.findContours doesn't think diagonal pixels are separate contours
just_cycles = erode(just_cycles, 2, 1)
# Use pcv.find_objects to turn plots of holes into countable contours
cycle_objects, cycle_hierarchies = find_objects(just_cycles, just_cycles)
# Count the number of holes
num_cycles = len(cycle_objects)
# Make debugging image
cycle_img = skel_img.copy()
cycle_img = dilate(cycle_img, params.line_thickness, 1)
cycle_img = cv2.cvtColor(cycle_img, cv2.COLOR_GRAY2RGB)
if num_cycles > 0:
rand_color = color_palette(num_cycles)
for i, cnt in enumerate(cycle_objects):
cv2.drawContours(cycle_img,
cycle_objects,
i,
rand_color[i],
params.line_thickness,
lineType=8,
hierarchy=cycle_hierarchies)
# Store Cycle Data
outputs.add_observation(variable='num_cycles',
trait='number of cycles',
method='plantcv.plantcv.morphology.check_cycles',
scale='none',
datatype=int,
value=num_cycles,
label='none')
# Reset debug mode
params.debug = debug
# Auto-increment device
params.device += 1
if params.debug == 'print':
print_image(
cycle_img,
os.path.join(params.debug_outdir,
str(params.device) + '_cycles.png'))
elif params.debug == 'plot':
plot_image(cycle_img)
return cycle_img
def get_coordinates(self):
if self.debug:
pcv.params.debug = 'print' # set debug mode
pcv.params.debug_outdir = './images/DracaenaVision/' # set output directory
blue_threshold = pcv.threshold.binary(gray_img=self.blue,
threshold=50,
max_value=255,
object_type='dark')
pcv.apply_mask(self.colour_image,
mask=blue_threshold,
mask_color='white')
# Calculate moments of binary image
moments = cv.moments(blue_threshold)
# Calculate x,y coordinate of center
self.centre_x = int(moments["m10"] / moments["m00"])
self.centre_y = int(moments["m01"] / moments["m00"])
# Put text and highlight the center
cv.circle(self.colour_image, (self.centre_x, self.centre_y), 5,
(255, 255, 255), -1)
cv.putText(self.colour_image, "Centre",
(self.centre_x - 25, self.centre_y - 25),
cv.FONT_HERSHEY_SIMPLEX, 0.5, (255, 255, 255), 2)
red_threshold = pcv.threshold.binary(gray_img=self.red,
threshold=70,
max_value=255,
object_type='light')
# Erode/Dilate the red threshold to remove noise
red_threshold = pcv.erode(red_threshold, ksize=5, i=1)
red_threshold = pcv.dilate(red_threshold, ksize=5, i=2)
# Create mask of area that is of interest
mask_pts = np.array(
[[576, 415], [800, 420], [1105, 630], [720, 685], [285, 590]],
np.int32).reshape((-1, 1, 2))
area_of_interest = np.zeros(self.colour_image.shape[:2], np.uint8)
cv.fillPoly(area_of_interest, [mask_pts], (255, 255, 255))
red_threshold_in_area_of_interest = pcv.logical_and(
area_of_interest, red_threshold)
pcv.apply_mask(img=self.colour_image,
mask=red_threshold_in_area_of_interest,
mask_color='black')
params = cv.SimpleBlobDetector_Params()
# Unused filters
params.filterByCircularity = False
params.filterByConvexity = False
params.filterByInertia = False
# Area filter
params.filterByArea = True
params.maxArea = 50000
params.minArea = 100
# Colour filter
params.filterByColor = True
params.blobColor = 255
# Misc options
params.minDistBetweenBlobs = 100
blob_detector = cv.SimpleBlobDetector_create(params)
keypoints = blob_detector.detect(red_threshold, mask=area_of_interest)
keypoints.sort(reverse=True, key=lambda kp: kp.size)
now = datetime.now().strftime('%d-%m %H %M %S')
im_with_keypoints = cv.drawKeypoints(
self.colour_image, keypoints, np.array([]), (0, 0, 255),
cv.DRAW_MATCHES_FLAGS_DRAW_RICH_KEYPOINTS)
self.detection_image = im_with_keypoints.copy()
if self.full_screen:
self.detection_image = cv.resize(self.detection_image,
(1920, 1080))
cv.imshow(self.window_name, self.detection_image)
cv.waitKey(1)
cv.polylines(im_with_keypoints, [mask_pts],
True, (0, 255, 0),
thickness=3)
cv.imwrite('./images/DracaenaVision/{} detections.png'.format(now),
im_with_keypoints)
coordinates = []
for keypoint in keypoints:
x = keypoint.pt[0]
y = keypoint.pt[1]
# For each X and Y determine depth (z-coordinate)
depth = -1
step_size_to_centre_x = (self.centre_x - x) / 100
step_size_to_centre_y = (self.centre_y - y) / 100
for i in range(21):
delta_x = step_size_to_centre_x * i
delta_y = step_size_to_centre_y * i
new_x = round(x + delta_x)
new_y = round(y + delta_y)
depth = self.depth_array[new_y, new_x]
if depth < 0.55:
print('Depth found with delta ({},{})'.format(
delta_x, delta_y))
print('Depth of: {} at {} X, {} Y'.format(
depth, new_x, new_y))
break
z = depth if depth < 0.55 or depth <= 0 else 0.55
# Determine the angle at which the tool should be held towards the plant
angle = 180 - math.degrees(
math.atan2(y - self.centre_y, x - self.centre_x))
coordinate = [x, y, z, angle]
coordinates.append(coordinate)
return coordinates
def main():
args = options() #create options object for argument parsing
device = 0 #set device
params.debug = args.debug #set debug
outfile = False
if args.writeimg:
outfile = os.path.join(args.outdir, os.path.basename(args.image)[:-4])
# In[114]:
img, path, filename = pcv.readimage(filename=args.image,
debug=args.debug) #read in image
background = pcv.transform.load_matrix(
args.npz) #read in background mask image for subtraction
# In[115]:
device, mask = pcv.naive_bayes_classifier(
img, args.pdf, device, args.debug) #naive bayes on image
#if args.writeimg:
# pcv.print_image(img=mask["94,104,47"], filename=outfile + "_nb_mask.png")
# In[116]:
new_mask = pcv.image_subtract(mask["94,104,47"],
background) #subtract background noise
# In[117]:
#image blurring using scipy median filter
blurred_img = ndimage.median_filter(new_mask, (7, 1))
blurred_img = ndimage.median_filter(blurred_img, (1, 7))
device, cleaned = pcv.fill(np.copy(blurred_img), np.copy(blurred_img), 50,
0, args.debug) #fill leftover noise
# In[118]:
#dilate and erode to repair plant breaks from background subtraction
device, cleaned_dilated = pcv.dilate(cleaned, 6, 1, 0)
device, cleaned = pcv.erode(cleaned_dilated, 6, 1, 0, args.debug)
# In[119]:
device, objects, obj_hierarchy = pcv.find_objects(
img, cleaned, device, debug=args.debug) #find objects using mask
if "TM015" in args.image:
h = 1620
elif "TM016" in args.image:
h = 1555
else:
h = 1320
roi_contour, roi_hierarchy = pcv.roi.rectangle(x=570,
y=0,
h=h,
w=1900 - 550,
img=img) #grab ROI
# In[120]:
#isolate plant objects within ROI
device, roi_objects, hierarchy, kept_mask, obj_area = pcv.roi_objects(
img,
'partial',
roi_contour,
roi_hierarchy,
objects,
obj_hierarchy,
device,
debug=args.debug)
#Analyze only images with plants present.
if roi_objects > 0:
# In[121]:
# Object combine kept objects
device, plant_contour, plant_mask = pcv.object_composition(
img=img,
contours=roi_objects,
hierarchy=hierarchy,
device=device,
debug=args.debug)
if args.writeimg:
pcv.print_image(img=plant_mask, filename=outfile + "_mask.png")
# In[122]:
# Find shape properties, output shape image (optional)
device, shape_header, shape_data, shape_img = pcv.analyze_object(
img=img,
imgname=args.image,
obj=plant_contour,
mask=plant_mask,
device=device,
debug=args.debug,
filename=outfile + ".png")
# In[123]:
if "TM015" in args.image:
line_position = 380
elif "TM016" in args.image:
line_position = 440
else:
line_position = 690
# Shape properties relative to user boundary line (optional)
device, boundary_header, boundary_data, boundary_img = pcv.analyze_bound_horizontal(
img=img,
obj=plant_contour,
mask=plant_mask,
line_position=line_position,
device=device,
debug=args.debug,
filename=outfile + ".png")
# In[124]:
# Determine color properties: Histograms, Color Slices and Pseudocolored Images,
# output color analyzed images (optional)
device, color_header, color_data, color_img = pcv.analyze_color(
img=img,
imgname=args.image,
mask=plant_mask,
bins=256,
device=device,
debug=args.debug,
hist_plot_type=None,
pseudo_channel="v",
pseudo_bkg="img",
resolution=300,
filename=outfile + ".png")
# In[55]:
# Output shape and color data
result = open(args.result, "a")
result.write('\t'.join(map(str, shape_header)) + "\n")
result.write('\t'.join(map(str, shape_data)) + "\n")
for row in shape_img:
result.write('\t'.join(map(str, row)) + "\n")
result.write('\t'.join(map(str, color_header)) + "\n")
result.write('\t'.join(map(str, color_data)) + "\n")
result.write('\t'.join(map(str, boundary_header)) + "\n")
result.write('\t'.join(map(str, boundary_data)) + "\n")
result.write('\t'.join(map(str, boundary_img)) + "\n")
for row in color_img:
result.write('\t'.join(map(str, row)) + "\n")
result.close()
def main():
# Get options
args = options()
if args.debug:
pcv.params.debug = args.debug # set debug mode
if args.debugdir:
pcv.params.debug_outdir = args.debugdir # set debug directory
os.makedirs(args.debugdir, exist_ok=True)
# pixel_resolution
# mm
# see pixel_resolution.xlsx for calibration curve for pixel to mm translation
pixelresolution = 0.052
# The result file should exist if plantcv-workflow.py was run
if os.path.exists(args.result):
# Open the result file
results = open(args.result, "r")
# The result file would have image metadata in it from plantcv-workflow.py, read it into memory
metadata = json.load(results)
# Close the file
results.close()
# Delete the file, we will create new ones
os.remove(args.result)
plantbarcode = metadata['metadata']['plantbarcode']['value']
print(plantbarcode,
metadata['metadata']['timestamp']['value'],
sep=' - ')
else:
# If the file did not exist (for testing), initialize metadata as an empty string
metadata = "{}"
regpat = re.compile(args.regex)
plantbarcode = re.search(regpat, args.image).groups()[0]
# read images and create mask
img, _, fn = pcv.readimage(args.image)
imagename = os.path.splitext(fn)[0]
# create mask
# taf=filters.try_all_threshold(s_img)
## remove background
s_img = pcv.rgb2gray_hsv(img, 's')
min_s = filters.threshold_minimum(s_img)
thresh_s = pcv.threshold.binary(s_img, min_s, 255, 'light')
rm_bkgrd = pcv.fill_holes(thresh_s)
## low greenness
thresh_s = pcv.threshold.binary(s_img, min_s + 15, 255, 'dark')
# taf = filters.try_all_threshold(s_img)
c = pcv.logical_xor(rm_bkgrd, thresh_s)
cinv = pcv.invert(c)
cinv_f = pcv.fill(cinv, 500)
cinv_f_c = pcv.closing(cinv_f, np.ones((5, 5)))
cinv_f_c_e = pcv.erode(cinv_f_c, 2, 1)
## high greenness
a_img = pcv.rgb2gray_lab(img, channel='a')
# taf = filters.try_all_threshold(a_img)
t_a = filters.threshold_isodata(a_img)
thresh_a = pcv.threshold.binary(a_img, t_a, 255, 'dark')
thresh_a = pcv.closing(thresh_a, np.ones((5, 5)))
thresh_a_f = pcv.fill(thresh_a, 500)
## combined mask
lor = pcv.logical_or(cinv_f_c_e, thresh_a_f)
close = pcv.closing(lor, np.ones((2, 2)))
fill = pcv.fill(close, 800)
erode = pcv.erode(fill, 3, 1)
fill2 = pcv.fill(erode, 1200)
# dilate = pcv.dilate(fill2,2,2)
mask = fill2
final_mask = np.zeros_like(mask)
# Compute greenness
# split color channels
b, g, r = cv2.split(img)
# print green intensity
# g_img = pcv.visualize.pseudocolor(g, cmap='Greens', background='white', min_value=0, max_value=255, mask=mask, axes=False)
# convert color channels to int16 so we can add them (values will be greater than 255 which is max of current uint8 format)
g = g.astype('uint16')
r = r.astype('uint16')
b = b.astype('uint16')
denom = g + r + b
# greenness index
out_flt = np.zeros_like(denom, dtype='float32')
# divide green by sum of channels to compute greenness index with values 0-1
gi = np.divide(g,
denom,
out=out_flt,
where=np.logical_and(denom != 0, mask > 0))
# find objects
c, h = pcv.find_objects(img, mask)
rc, rh = pcv.roi.multi(img, coord=[(1300, 900), (1300, 2400)], radius=350)
# Turn off debug temporarily, otherwise there will be a lot of plots
pcv.params.debug = None
# Loop over each region of interest
i = 0
rc_i = rc[i]
for i, rc_i in enumerate(rc):
rh_i = rh[i]
# Add ROI number to output. Before roi_objects so result has NA if no object.
pcv.outputs.add_observation(variable='roi',
trait='roi',
method='roi',
scale='int',
datatype=int,
value=i,
label='#')
roi_obj, hierarchy_obj, submask, obj_area = pcv.roi_objects(
img,
roi_contour=rc_i,
roi_hierarchy=rh_i,
object_contour=c,
obj_hierarchy=h,
roi_type='partial')
if obj_area == 0:
print('\t!!! No object found in ROI', str(i))
pcv.outputs.add_observation(
variable='plantarea',
trait='plant area in sq mm',
method='observations.area*pixelresolution^2',
scale=pixelresolution,
datatype="<class 'float'>",
value=0,
label='sq mm')
else:
# Combine multiple objects
# ple plant objects within an roi together
plant_object, plant_mask = pcv.object_composition(
img=img, contours=roi_obj, hierarchy=hierarchy_obj)
final_mask = pcv.image_add(final_mask, plant_mask)
# Save greenness for individual ROI
grnindex = np.mean(gi[np.where(plant_mask > 0)])
pcv.outputs.add_observation(
variable='greenness_index',
trait='mean normalized greenness index',
method='g/sum(b+g+r)',
scale='[0,1]',
datatype="<class 'float'>",
value=float(grnindex),
label='/1')
# Analyze all colors
hist = pcv.analyze_color(img, plant_mask, 'all')
# Analyze the shape of the current plant
shape_img = pcv.analyze_object(img, plant_object, plant_mask)
plant_area = pcv.outputs.observations['area'][
'value'] * pixelresolution**2
pcv.outputs.add_observation(
variable='plantarea',
trait='plant area in sq mm',
method='observations.area*pixelresolution^2',
scale=pixelresolution,
datatype="<class 'float'>",
value=plant_area,
label='sq mm')
# end if-else
# At this point we have observations for one plant
# We can write these out to a unique results file
# Here I will name the results file with the ROI ID combined with the original result filename
basename, ext = os.path.splitext(args.result)
filename = basename + "-roi" + str(i) + ext
# Save the existing metadata to the new file
with open(filename, "w") as r:
json.dump(metadata, r)
pcv.print_results(filename=filename)
# The results are saved, now clear out the observations so the next loop adds new ones for the next plant
pcv.outputs.clear()
if args.writeimg and obj_area != 0:
imgdir = os.path.join(args.outdir, 'shape_images', plantbarcode)
os.makedirs(imgdir, exist_ok=True)
pcv.print_image(
shape_img,
os.path.join(imgdir,
imagename + '-roi' + str(i) + '-shape.png'))
imgdir = os.path.join(args.outdir, 'colorhist_images',
plantbarcode)
os.makedirs(imgdir, exist_ok=True)
pcv.print_image(
hist,
os.path.join(imgdir,
imagename + '-roi' + str(i) + '-colorhist.png'))
# end roi loop
if args.writeimg:
# save grnness image of entire tray
imgdir = os.path.join(args.outdir, 'pseudocolor_images', plantbarcode)
os.makedirs(imgdir, exist_ok=True)
gi_img = pcv.visualize.pseudocolor(gi,
obj=None,
mask=final_mask,
cmap='viridis',
axes=False,
min_value=0.3,
max_value=0.6,
background='black',
obj_padding=0)
gi_img = add_scalebar(gi_img,
pixelresolution=pixelresolution,
barwidth=20,
barlocation='lower left')
gi_img.set_size_inches(6, 6, forward=False)
gi_img.savefig(os.path.join(imgdir, imagename + '-greenness.png'),
bbox_inches='tight')
gi_img.clf()
############################################
###### Perform Calculations ##########
############################################
# combine leaf and labels
mask_image_label, path, filename = pcv.readimage(
'1_naive_bayes_labels_mask.jpg')
mask_image = mask_image_plant + mask_image_label
device, mask_image = pcv.rgb2gray_lab(mask_image, 'l', device)
#clean the mask up
device, img_binary = pcv.binary_threshold(mask_image, 50, 255, 'light', device)
pcv.print_image(img_binary, 'img_binary.tif')
device, blur_img = pcv.erode(
img_binary, 3, 1, device, debug='print'
) # Erode to remove soil and Dilate so that you don't lose leaves (just in case)
mask = np.copy(blur_img)
device, fill_image = pcv.fill(blur_img, mask, 100, device)
pcv.print_image(fill_image, 'fill_image.tif')
device, binary_image = pcv.median_blur(fill_image, 1, device)
pcv.print_image(binary_image, 'binary_image.tif')
device, masked_image = device, dilate_image = pcv.dilate(
fill_image, 3, 3, device)
############################################
########### Create Output #############
############################################
#Print grid of images for QC
def main_side():
# Setting "args"
# Get options
pcv.params.debug = args.debug #set debug mode
pcv.params.debug_outdir = args.outdir #set output directory
# Read image (readimage mode defaults to native but if image is RGBA then specify mode='rgb')
# Inputs:
# filename - Image file to be read in
# mode - Return mode of image; either 'native' (default), 'rgb', 'gray', or 'csv'
filename = args.image
img = cv2.imread(args.image, flags=0)
#img = pcv.invert(img)
path, img_name = os.path.split(args.image)
img_bkgrd = cv2.imread("background.png", flags=0)
#print(img)
#print(img_bkgrd)
bkg_sub_img = pcv.image_subtract(img_bkgrd, img)
bkg_sub_thres_img, masked_img = pcv.threshold.custom_range(
rgb_img=bkg_sub_img,
lower_thresh=[50],
upper_thresh=[255],
channel='gray')
# Laplace filtering (identify edges based on 2nd derivative)
# Inputs:
# gray_img - Grayscale image data
# ksize - Aperture size used to calculate the second derivative filter,
# specifies the size of the kernel (must be an odd integer)
# scale - Scaling factor applied (multiplied) to computed Laplacian values
# (scale = 1 is unscaled)
lp_img = pcv.laplace_filter(gray_img=img, ksize=1, scale=1)
# Plot histogram of grayscale values
pcv.visualize.histogram(gray_img=lp_img)
# Lapacian image sharpening, this step will enhance the darkness of the edges detected
lp_shrp_img = pcv.image_subtract(gray_img1=img, gray_img2=lp_img)
# Plot histogram of grayscale values, this helps to determine thresholding value
pcv.visualize.histogram(gray_img=lp_shrp_img)
# Sobel filtering
# 1st derivative sobel filtering along horizontal axis, kernel = 1)
# Inputs:
# gray_img - Grayscale image data
# dx - Derivative of x to analyze
# dy - Derivative of y to analyze
# ksize - Aperture size used to calculate 2nd derivative, specifies the size of the kernel and must be an odd integer
# NOTE: Aperture size must be greater than the largest derivative (ksize > dx & ksize > dy)
sbx_img = pcv.sobel_filter(gray_img=img, dx=1, dy=0, ksize=1)
# 1st derivative sobel filtering along vertical axis, kernel = 1)
sby_img = pcv.sobel_filter(gray_img=img, dx=0, dy=1, ksize=1)
# Combine the effects of both x and y filters through matrix addition
# This will capture edges identified within each plane and emphasize edges found in both images
# Inputs:
# gray_img1 - Grayscale image data to be added to gray_img2
# gray_img2 - Grayscale image data to be added to gray_img1
sb_img = pcv.image_add(gray_img1=sbx_img, gray_img2=sby_img)
# Use a lowpass (blurring) filter to smooth sobel image
# Inputs:
# gray_img - Grayscale image data
# ksize - Kernel size (integer or tuple), (ksize, ksize) box if integer input,
# (n, m) box if tuple input
mblur_img = pcv.median_blur(gray_img=sb_img, ksize=1)
# Inputs:
# gray_img - Grayscale image data
mblur_invert_img = pcv.invert(gray_img=mblur_img)
# combine the smoothed sobel image with the laplacian sharpened image
# combines the best features of both methods as described in "Digital Image Processing" by Gonzalez and Woods pg. 169
edge_shrp_img = pcv.image_add(gray_img1=mblur_invert_img,
gray_img2=lp_shrp_img)
# Perform thresholding to generate a binary image
tr_es_img = pcv.threshold.binary(gray_img=edge_shrp_img,
threshold=145,
max_value=255,
object_type='dark')
# Do erosion with a 3x3 kernel (ksize=3)
# Inputs:
# gray_img - Grayscale (usually binary) image data
# ksize - The size used to build a ksize x ksize
# matrix using np.ones. Must be greater than 1 to have an effect
# i - An integer for the number of iterations
e1_img = pcv.erode(gray_img=tr_es_img, ksize=3, i=1)
# Bring the two object identification approaches together.
# Using a logical OR combine object identified by background subtraction and the object identified by derivative filter.
# Inputs:
# bin_img1 - Binary image data to be compared in bin_img2
# bin_img2 - Binary image data to be compared in bin_img1
comb_img = pcv.logical_or(bin_img1=e1_img, bin_img2=bkg_sub_thres_img)
# Get masked image, Essentially identify pixels corresponding to plant and keep those.
# Inputs:
# rgb_img - RGB image data
# mask - Binary mask image data
# mask_color - 'black' or 'white'
masked_erd = pcv.apply_mask(rgb_img=img, mask=comb_img, mask_color='black')
# Need to remove the edges of the image, we did that by generating a set of rectangles to mask the edges
# img is (1280 X 960)
# mask for the bottom of the image
# Inputs:
# img - RGB or grayscale image data
# p1 - Point at the top left corner of the rectangle (tuple)
# p2 - Point at the bottom right corner of the rectangle (tuple)
# color 'black' (default), 'gray', or 'white'
#
masked1, box1_img, rect_contour1, hierarchy1 = pcv.rectangle_mask(img=img,
p1=(500,
875),
p2=(720,
960))
# mask the edges
masked2, box2_img, rect_contour2, hierarchy2 = pcv.rectangle_mask(img=img,
p1=(1,
1),
p2=(1279,
959))
bx12_img = pcv.logical_or(bin_img1=box1_img, bin_img2=box2_img)
inv_bx1234_img = bx12_img # we dont invert
inv_bx1234_img = bx12_img
#inv_bx1234_img = pcv.invert(gray_img=bx12_img)
edge_masked_img = pcv.apply_mask(rgb_img=masked_erd,
mask=inv_bx1234_img,
mask_color='black')
#print("here we create a mask")
mask, masked = pcv.threshold.custom_range(rgb_img=edge_masked_img,
lower_thresh=[25],
upper_thresh=[175],
channel='gray')
masked = pcv.apply_mask(rgb_img=masked, mask=mask, mask_color='white')
#print("end")
# Identify objects
# Inputs:
# img - RGB or grayscale image data for plotting
# mask - Binary mask used for detecting contours
id_objects, obj_hierarchy = pcv.find_objects(img=edge_masked_img,
mask=mask)
# Define ROI
# Inputs:
# img - RGB or grayscale image to plot the ROI on
# x - The x-coordinate of the upper left corner of the rectangle
# y - The y-coordinate of the upper left corner of the rectangle
# h - The height of the rectangle
# w - The width of the rectangle
roi1, roi_hierarchy = pcv.roi.rectangle(img=edge_masked_img,
x=100,
y=100,
h=800,
w=1000)
# Decide which objects to keep
# Inputs:
# img = img to display kept objects
# roi_contour = contour of roi, output from any ROI function
# roi_hierarchy = contour of roi, output from any ROI function
# object_contour = contours of objects, output from pcv.find_objects function
# obj_hierarchy = hierarchy of objects, output from pcv.find_objects function
# roi_type = 'partial' (default, for partially inside), 'cutto', or
# 'largest' (keep only largest contour)
with HiddenPrints():
roi_objects, hierarchy5, kept_mask, obj_area = pcv.roi_objects(
img=edge_masked_img,
roi_contour=roi1,
roi_hierarchy=roi_hierarchy,
object_contour=id_objects,
obj_hierarchy=obj_hierarchy,
roi_type='largest')
rgb_img = cv2.cvtColor(img, cv2.COLOR_GRAY2RGB)
# Inputs:
# img - RGB or grayscale image data for plotting
# contours - Contour list
# hierarchy - Contour hierarchy array
o, m = pcv.object_composition(img=rgb_img,
contours=roi_objects,
hierarchy=hierarchy5)
### Analysis ###
outfile = False
if args.writeimg == True:
outfile = args.outdir + "/" + filename
# Perform signal analysis
# Inputs:
# img - RGB or grayscale image data
# obj- Single or grouped contour object
# mask - Binary image mask to use as mask for moments analysis
shape_img = pcv.analyze_object(img=img, obj=o, mask=m)
new_im = Image.fromarray(shape_img)
new_im.save("output//" + args.filename + "shape_img_side.png")
# Inputs:
# gray_img - 8 or 16-bit grayscale image data
# mask - Binary mask made from selected contours
# bins - Number of classes to divide the spectrum into
# histplot - If True, plots the histogram of intensity values
nir_hist = pcv.analyze_nir_intensity(gray_img=img,
mask=kept_mask,
bins=256,
histplot=True)
# Pseudocolor the grayscale image to a colormap
# Inputs:
# gray_img - Grayscale image data
# obj - Single or grouped contour object (optional), if provided the pseudocolored image gets cropped down to the region of interest.
# mask - Binary mask (optional)
# background - Background color/type. Options are "image" (gray_img), "white", or "black". A mask must be supplied.
# cmap - Colormap
# min_value - Minimum value for range of interest
# max_value - Maximum value for range of interest
# dpi - Dots per inch for image if printed out (optional, if dpi=None then the default is set to 100 dpi).
# axes - If False then the title, x-axis, and y-axis won't be displayed (default axes=True).
# colorbar - If False then the colorbar won't be displayed (default colorbar=True)
pseudocolored_img = pcv.visualize.pseudocolor(gray_img=img,
mask=kept_mask,
cmap='viridis')
# Perform shape analysis
# Inputs:
# img - RGB or grayscale image data
# obj- Single or grouped contour object
# mask - Binary image mask to use as mask for moments analysis
shape_imgs = pcv.analyze_object(img=rgb_img, obj=o, mask=m)
# Write shape and nir data to results file
pcv.print_results(filename=args.result)
import cv2
pcv.params.debug = 'plot'
plt.rcParams['figure.figsize'] = [12, 12]
visfn = "data/vis/A6-GoldStandard2_RGB-20190801T043947-VIS0-0.png"
psIIfn = "data/psII/A6-GoldStandard2_PSII-20190801T003230-PSII0-2.png"
visimg, bn, fn = pcv.readimage(visfn)
psIIimg, bn, fn = pcv.readimage(psIIfn)
vis_x = visimg.shape[1]
vis_y = visimg.shape[0]
psII_x = psIIimg.shape[1]
psII_y = psIIimg.shape[0]
masko = pcv.threshold.otsu(psIIimg, 255, 'light')
mask = pcv.erode(masko, 2, 2)
final_mask = mask
mask_shift_x = pcv.shift_img(final_mask, 14, 'left')
mask_shift_y = pcv.shift_img(mask_shift_x, 3, 'top')
# vis_mask = pcv.resize(final_mask, resize_x = vis_x/psII_x, resize_y=vis_y/psII_y)
vis_mask2 = cv2.resize(mask_shift_y, (vis_x, vis_y),
interpolation=cv2.INTER_CUBIC)
vis_masked = pcv.apply_mask(visimg, vis_mask2, 'black')
# vs_ws=pcv.watershed_segmentation(visimg,vis_mask2)
def check_cycles(skel_img):
""" Check for cycles in a skeleton image
Inputs:
skel_img = Skeletonized image
Returns:
cycle_img = Image with cycles identified
:param skel_img: numpy.ndarray
:return cycle_img: numpy.ndarray
"""
# Store debug
debug = params.debug
params.debug = None
# Create the mask needed for cv2.floodFill, must be larger than the image
h, w = skel_img.shape[:2]
mask = np.zeros((h + 2, w + 2), np.uint8)
# Copy the skeleton since cv2.floodFill will draw on it
skel_copy = skel_img.copy()
cv2.floodFill(skel_copy, mask=mask, seedPoint=(0, 0), newVal=255)
# Invert so the holes are white and background black
just_cycles = cv2.bitwise_not(skel_copy)
# Erode slightly so that cv2.findContours doesn't think diagonal pixels are separate contours
just_cycles = erode(just_cycles, 2, 1)
# Use pcv.find_objects to turn plots of holes into countable contours
cycle_objects, cycle_hierarchies = find_objects(just_cycles, just_cycles)
# Count the number of holes
num_cycles = len(cycle_objects)
# Make debugging image
cycle_img = skel_img.copy()
cycle_img = dilate(cycle_img, params.line_thickness, 1)
cycle_img = cv2.cvtColor(cycle_img, cv2.COLOR_GRAY2RGB)
rand_color = color_palette(num_cycles)
for i, cnt in enumerate(cycle_objects):
cv2.drawContours(cycle_img, cycle_objects, i, rand_color[i], params.line_thickness, lineType=8,
hierarchy=cycle_hierarchies)
# Store Cycle Data
outputs.add_observation(variable='num_cycles', trait='number of cycles',
method='plantcv.plantcv.morphology.check_cycles', scale='none', datatype=int,
value=num_cycles, label='none')
# Reset debug mode
params.debug = debug
# Auto-increment device
params.device += 1
if params.debug == 'print':
print_image(cycle_img, os.path.join(params.debug_outdir, str(params.device) + '_cycles.png'))
elif params.debug == 'plot':
plot_image(cycle_img)
return cycle_img