def load_training_samples():
""" Load the training samples. """
train_image_path = Settings.get("TRAIN_IMAGE_PATH")
label_path = Settings.get("LABEL_PATH")
data_frame = pandas.read_csv(label_path)
samples: List[Sample] = []
for index, row in data_frame.iterrows():
sample = Sample(image_id=row["Image"], image_dir=train_image_path, label=row["Id"])
samples.append(sample)
return samples
def reset(self):
"""
Resets settings to factory settings (defaults)
:return: None
"""
self.settings = Settings()
self.init_settings()
def apply(self):
"""
Gather data into settings variable and set it as the result
:return: None
"""
self.__result = Settings(
float(self.harris_score_threshold_spinbox.get()),
float(self.harris_free_parameter_spinbox.get()),
int(float(self.neighborhood_size_spinbox.get())), self.mark_color,
int(float(self.canny_min_threshold_entry.get())),
int(float(self.canny_max_threshold_entry.get())),
bool(self.rect_mark_check_var.get()),
int(float(self.dilate_size_spinbox.get())), self.quality_var.get())
def load_testing_samples():
""" Load the training samples. """
test_image_path = Settings.get("TEST_IMAGE_PATH")
samples: List[Sample] = []
files = os.listdir(test_image_path)
for file_name in files:
if ".jpg" not in file_name and ".png" not in file_name:
continue
sample = Sample(image_id=file_name, image_dir=test_image_path)
samples.append(sample)
return samples
class Manager():
__metaclass__ = ABCMeta
conf = Settings().settings
def __init__(self, manager, data=None):
self.manager = manager
self.data = data
@abstractmethod
def parse(self):
pass
def needIndex():
if Manager.conf['info_license'] == True or \
Manager.conf['info_author'] == True or \
Manager.conf['info_homepage_url'] == True or \
Manager.conf['info_code_url'] == True:
return True
return False
def needLicense():
if Manager.conf['info_license'] == True:
return True
return False
def needAuthor():
if Manager.conf['info_author'] == True:
return True
return False
def needHpUrl():
if Manager.conf['info_homepage_url'] == True:
return True
return False
def needCodeUrl():
if Manager.conf['info_code_url'] == True:
return True
return False
def _reallyReload():
Settings.reloadStage1()
Settings.reloadStage2()
default=False, help="Force overwrite of existing config when creating config.")
# CONSIDER: this could easily support multiple config files I guess
# but changing Settings to support this would be kind of intense I think.
parser.add_argument('config', nargs="?", metavar="CONFIGFILE", default="BurlyBot.json")
args = parser.parse_args()
# create-config
if args.createconfig:
if not args.config: args.config = "BurlyBot.json"
print "Creating configuration..."
if exists(args.config) and not args.force:
print "Error: NEWCONFIGFILE (%s) exists. Use --force (-f) to force overwrite. Bailing." % args.config
exit(1)
Settings.configfile = args.config
Settings.saveOptions()
print "Done."
exit(0)
if args.config and exists(args.config):
Settings.configfile = args.config
else:
print "Error: Settings file (%s) not found." % args.config
exit(2)
try:
Settings.load()
except ConfigException as e:
print "Error:", e
exit(2)
Settings.initialize(logger=templog)
parser.add_argument('config',
nargs="?",
metavar="CONFIGFILE",
default="BurlyBot.json")
args = parser.parse_args()
# create-config
if args.createconfig:
if not args.config: args.config = "BurlyBot.json"
print "Creating configuration..."
if exists(args.config) and not args.force:
print "Error: NEWCONFIGFILE (%s) exists. Use --force (-f) to force overwrite. Bailing." % args.config
exit(1)
Settings.configfile = args.config
Settings.saveOptions()
print "Done."
exit(0)
if args.config and exists(args.config):
Settings.configfile = args.config
else:
print "Error: Settings file (%s) not found." % args.config
exit(2)
try:
Settings.load()
except ConfigException as e:
print "Error:", e
exit(2)
Settings.initialize(logger=templog)
def _reallyReload(): Settings.reloadStage1() Settings.reloadStage2()
parser.add_option("-c", "--choice", dest="choice", default = False, type=int,
help="0 to build trajectories, 1 to compute xml files from trajectories, 2 to compute features from existing trajectories")
parser.add_option("-s", "--simulated", dest="simulated", default = 0, type=int,
help="Use of simulated trajectories or no")
parser.add_option("--cecog_file", dest="cecog_file", default = False, type=int,
help="True for new type, False for old")
(options, args) = parser.parse_args()
if options.well==None:
print "You need to specify which well to treat. Pgm exiting"
sys.exit()
if type(options.choice)!=bool: options.choice=int(options.choice)
settings = Settings(options.settings_file, globals())
if not os.path.isdir(settings.outputFolder):
os.mkdir(settings.outputFolder)
outputFolder = os.path.join(settings.outputFolder, options.plate)
if not os.path.isdir(outputFolder):
os.mkdir(outputFolder)
fi=settings.traj_filename.format(options.well)
fi_trajfeatures = settings.feature_filename.format(options.well.split('.')[0])
if options.simulated:
settings.training=True
outputFolder = os.path.join('../resultData/simulated_traj/simres/plates', options.plate)
fi='{}--{}.pickle'.format(options.plate, options.well)
def corners_and_line_intersection_detector(img,
console_consumer=None,
progress_consumer=None,
is_using_canny=False,
is_applying_gauss=False,
settings=Settings()):
"""
Main entry point to the Harris Detector algorithm
We call this method from main frame in order to open an image, detect corners in it, and return both the image and
a copy of the image with the corners marked out
:param img: Image to work on
:param console_consumer: A lambda / function to handle status updates
:param progress_consumer: A lambda / function to handle progress updates
:param is_using_canny: Whether we should detect edges using Canny Edge Detection before Harris Detector, or not
:param is_applying_gauss: Whether to apply Gaussian Blur before executing Harris Corner Detector
:param settings: Settings to know how to execute the algorithm
:return: image, image_with_marks
"""
start = timer()
progress = do_progress(progress_consumer, 0, 5)
success = False
if img is None:
log(console_consumer, 'Specified image was None')
image = None
else:
image = img.copy()
if is_applying_gauss:
log(console_consumer, 'Applying Gaussian Blur...')
image = apply_gaussian_blur(
image) # yields either 3D or 2D, depends on the image
progress = do_progress(progress_consumer, progress, 10)
if is_using_canny:
log(console_consumer, 'Running Canny Edge Detection...')
image = cv2.Canny(image, settings.canny_min_thresh,
settings.canny_max_thresh) # yields 2D
progress = do_progress(progress_consumer, progress, 10)
log(console_consumer, 'Detecting corners...')
harris_scores = find_harris_corners(image,
settings.harris_free_parameter,
settings.neighborhood_size,
console_consumer)
progress = do_progress(progress_consumer, progress, 10)
if harris_scores is None:
log(console_consumer,
'Error has occurred while detecting corners. Result was None')
image = None
else:
success = True
image = img.copy()
if image.ndim == 2:
image = cv2.cvtColor(image, cv2.COLOR_GRAY2BGR)
# Mark the corners
log(console_consumer, 'Marking detected interest points...')
if settings.corners_quality == HIGH_QUALITY:
helper_image = enhance_corners_accuracy(
harris_scores, image, settings)
progress = do_progress(progress_consumer, progress, 10)
mark_corners(image, helper_image, settings, progress_consumer,
progress)
else:
# Faster marking, low quality
if settings.is_using_rect_mark:
mark_corners_with_rect(image, harris_scores, settings,
progress_consumer, progress)
else:
# Mark with dots
harris_scores = cv2.dilate(harris_scores, None)
image[harris_scores > settings.harris_score_threshold *
harris_scores.max()] = settings.corners_color[::-1]
time_took = timer() - start
do_progress(progress_consumer, 0, 100)
log(console_consumer,
'Corners detection ended in %.2f seconds. Result:' % time_took,
'Success' if success else 'Fail')
return img, image
from util.settings import Settings
from manager import classify_manager
from output import classify_output
from github import Github
def get_manager_from_path(path):
return path.split('/')[-1].replace('.', '_').lower()
args = argparser.parse()
library_info_list = []
manager_files = []
timestamp = datetime.now().strftime("%Y/%m/%d %H:%M:%S")
conf = Settings().settings
for repo in conf['target_repositories']:
git = Github(args.token)
manager_paths = git.search_files(repo, conf['target_manager'])
for path in manager_paths:
manager_files.append({'path': path, 'data': git.get_files(repo, path)})
for file in manager_files:
print('processing: ' + file['path'] + ' ...')
manager = classify_manager.new_manager(
get_manager_from_path(file['path']), file['data'])
if not manager:
continue
lib_list = manager.parse()
def __init__(self, master):
tk.Frame.__init__(self, master)
master.config(background=ctl.BACKGROUND_COLOR)
toplevel = self.winfo_toplevel()
toplevel.title('Corners Detector')
toplevel.iconbitmap(os.path.abspath(os.path.join('resource', 'corners-icon.ico')))
# Declare all of the instance attributes here, and initialize them in separate methods for code separation
# Keep them as attributes so we will not have issues with garbage collection. Especially with the PhotoImage
self.settings = Settings().load()
self.__title_frame = None # tk.Frame
self.__action_frame = None # tk.Frame
self.__status_frame = None # tk.Frame
self.__figure_frame = None # tk.Frame
self.__title = None # tk.Label
self.__status_bar = None # tk.Label
self.__progress_bar = None # tk.Progressbar
self.__settings_button = None # tk.Button
self.__go_button = None # tk.Button
self.__open_file_button = None # tk.Button
self.__popup_image_button = None # tk.Button
self.__settings_button_tooltip = None # view.tooltip.Tooltip
self.__go_button_tooltip = None # view.tooltip.Tooltip
self.__open_file_button_tooltip = None # view.tooltip.Tooltip
self.__popup_image_button_tooltip = None # view.tooltip.Tooltip
self.__file_path_entry_tooltip = None # view.tooltip.Tooltip
self.__canny_check_var = None # tk.IntVar - to hold the value of the canny edge checkbox
self.__canny_checkbutton = None # tk.Checkbutton
self.__gauss_check_var = None # tk.IntVar - to hold the value of the gaussian checkbox
self.__gauss_checkbutton = None # tk.Checkbutton
self.__file_path_entry = None # tk.Entry
self.__magnifying_icon = None # tk.PhotoImage
self.__play_icon = None # tk.PhotoImage
self.__popout_icon = None # tk.PhotoImage
self.__settings_icon = None # tk.PhotoImage
self.__figure = None # A reference to pyplot figure, so we can destroy it when there is a new input
self.__style = None # tk.ttk.Style
self.__running = False # Indication of when we wait for the Harris Detector worker to finish
self.__image = None # A reference to the input image. It is being set by the Harris Detector action
self.__processed_image = None # A reference to the processed image (outcome). It is being set by the Harris Detector action
self.__error = False # Indication for a failure during Harris Detector algorithm
self.progress_text_format = '{0}%'
self.__style = Style(master)
self.__style.theme_use('clam')
self.__style.configure("Horizontal.TProgressbar", foreground='#E0E0E0', background=ctl.ACCEPT_COLOR,
troughcolor=ctl.BACKGROUND_TOOLTIP_COLOR, bordercolor=ctl.ACCEPT_COLOR,
lightcolor=ctl.ACCEPT_COLOR, darkcolor=ctl.ACCEPT_COLOR)
self.__style.configure('TButton', font=ctl.FONT_BUTTON, borderwidth='1', background=ctl.BACKGROUND_COLOR, relief='FLAT')
self.__style.configure('TLabel', font=ctl.FONT_REGULAR)
self.__style.map('TButton', background=[('active', '!disabled', '#4E6067')])
self.__style.map('TCombobox', foreground=[('readonly', '!disabled', ctl.FOREGROUND_EDITOR_COLOR)])
self.__style.map('TCombobox', background=[('readonly', '!disabled', ctl.BACKGROUND_EDITOR_COLOR)])
self.__style.map('TCombobox', fieldbackground=[('readonly', '!disabled', ctl.BACKGROUND_EDITOR_COLOR)])
self.__style.map('TCombobox', lightcolor=[('readonly', '!disabled', 'black')])
self.__style.map('TCombobox', darkcolor=[('readonly', '!disabled', 'black')])
self.__style.map('TCombobox', bordercolor=[('readonly', '!disabled', '#E0E0E0')])
master.option_add("*TCombobox*Listbox*Background", ctl.BACKGROUND_EDITOR_COLOR)
master.option_add("*TCombobox*Listbox*Foreground", ctl.FOREGROUND_EDITOR_COLOR)
self.create_title_section(master)
self.create_action_section(master)
self.create_status_section(master)
self.create_workarea_section(master)
master.bind('<Return>', self.on_enter_pressed)
master.geometry("800x600")
class MainDialog(tk.Frame):
def __init__(self, master):
tk.Frame.__init__(self, master)
master.config(background=ctl.BACKGROUND_COLOR)
toplevel = self.winfo_toplevel()
toplevel.title('Corners Detector')
toplevel.iconbitmap(os.path.abspath(os.path.join('resource', 'corners-icon.ico')))
# Declare all of the instance attributes here, and initialize them in separate methods for code separation
# Keep them as attributes so we will not have issues with garbage collection. Especially with the PhotoImage
self.settings = Settings().load()
self.__title_frame = None # tk.Frame
self.__action_frame = None # tk.Frame
self.__status_frame = None # tk.Frame
self.__figure_frame = None # tk.Frame
self.__title = None # tk.Label
self.__status_bar = None # tk.Label
self.__progress_bar = None # tk.Progressbar
self.__settings_button = None # tk.Button
self.__go_button = None # tk.Button
self.__open_file_button = None # tk.Button
self.__popup_image_button = None # tk.Button
self.__settings_button_tooltip = None # view.tooltip.Tooltip
self.__go_button_tooltip = None # view.tooltip.Tooltip
self.__open_file_button_tooltip = None # view.tooltip.Tooltip
self.__popup_image_button_tooltip = None # view.tooltip.Tooltip
self.__file_path_entry_tooltip = None # view.tooltip.Tooltip
self.__canny_check_var = None # tk.IntVar - to hold the value of the canny edge checkbox
self.__canny_checkbutton = None # tk.Checkbutton
self.__gauss_check_var = None # tk.IntVar - to hold the value of the gaussian checkbox
self.__gauss_checkbutton = None # tk.Checkbutton
self.__file_path_entry = None # tk.Entry
self.__magnifying_icon = None # tk.PhotoImage
self.__play_icon = None # tk.PhotoImage
self.__popout_icon = None # tk.PhotoImage
self.__settings_icon = None # tk.PhotoImage
self.__figure = None # A reference to pyplot figure, so we can destroy it when there is a new input
self.__style = None # tk.ttk.Style
self.__running = False # Indication of when we wait for the Harris Detector worker to finish
self.__image = None # A reference to the input image. It is being set by the Harris Detector action
self.__processed_image = None # A reference to the processed image (outcome). It is being set by the Harris Detector action
self.__error = False # Indication for a failure during Harris Detector algorithm
self.progress_text_format = '{0}%'
self.__style = Style(master)
self.__style.theme_use('clam')
self.__style.configure("Horizontal.TProgressbar", foreground='#E0E0E0', background=ctl.ACCEPT_COLOR,
troughcolor=ctl.BACKGROUND_TOOLTIP_COLOR, bordercolor=ctl.ACCEPT_COLOR,
lightcolor=ctl.ACCEPT_COLOR, darkcolor=ctl.ACCEPT_COLOR)
self.__style.configure('TButton', font=ctl.FONT_BUTTON, borderwidth='1', background=ctl.BACKGROUND_COLOR, relief='FLAT')
self.__style.configure('TLabel', font=ctl.FONT_REGULAR)
self.__style.map('TButton', background=[('active', '!disabled', '#4E6067')])
self.__style.map('TCombobox', foreground=[('readonly', '!disabled', ctl.FOREGROUND_EDITOR_COLOR)])
self.__style.map('TCombobox', background=[('readonly', '!disabled', ctl.BACKGROUND_EDITOR_COLOR)])
self.__style.map('TCombobox', fieldbackground=[('readonly', '!disabled', ctl.BACKGROUND_EDITOR_COLOR)])
self.__style.map('TCombobox', lightcolor=[('readonly', '!disabled', 'black')])
self.__style.map('TCombobox', darkcolor=[('readonly', '!disabled', 'black')])
self.__style.map('TCombobox', bordercolor=[('readonly', '!disabled', '#E0E0E0')])
master.option_add("*TCombobox*Listbox*Background", ctl.BACKGROUND_EDITOR_COLOR)
master.option_add("*TCombobox*Listbox*Foreground", ctl.FOREGROUND_EDITOR_COLOR)
self.create_title_section(master)
self.create_action_section(master)
self.create_status_section(master)
self.create_workarea_section(master)
master.bind('<Return>', self.on_enter_pressed)
master.geometry("800x600")
def create_title_section(self, master):
"""
Title area contains a frame with a label in it.
:param master: Master dialog to add the frame to
:return: None
"""
self.__title_frame = ctl.create_frame(master, tk.X)
self.__title = ctl.create_title(self.__title_frame, 'Corners Detector')
self.__settings_icon = tk.PhotoImage(file=os.path.abspath(os.path.join('resource', 'settings-icon.png')))
self.__settings_button = ttk.Button(master=self.__title_frame, image=self.__settings_icon, command=self.show_settings_dialog,
style='TButton')
self.__settings_button_tooltip = Tooltip(self.__settings_button, 'Settings')
# First display the button and only then the label, so they will sit in one row
self.__settings_button.pack(side=tk.RIGHT)
self.__title.pack(fill=tk.X)
def create_action_section(self, master):
"""
Action area contains a frame with an entry (text edit for image path), open file dialog button and Go button.
Open file dialog button will display an open file dialog to select image
Go button will execute Harris Detector action on the selected image
:param master: Master dialog to add the frame to
:return: None
"""
self.__action_frame = ctl.create_frame(master, tk.X)
self.__magnifying_icon = tk.PhotoImage(file=os.path.abspath(os.path.join('resource', 'magnifying-icon.png')))
self.__file_path_entry = ctl.create_entry(self.__action_frame)
self.__file_path_entry.pack(fill=tk.BOTH, side=tk.LEFT, expand=True)
self.__file_path_entry_tooltip = Tooltip(self.__file_path_entry, 'Click the magnifying icon in order to open an image')
self.__open_file_button = ttk.Button(master=self.__action_frame, image=self.__magnifying_icon,
command=self.open_file_action, style='TButton', width=4)
self.__open_file_button.pack(side=tk.LEFT)
self.__open_file_button_tooltip = Tooltip(self.__open_file_button, 'Open image or video')
ctl.create_pad(self.__action_frame, tk.LEFT)
self.__play_icon = tk.PhotoImage(file=os.path.abspath(os.path.join('resource', 'play-icon.png')))
self.__go_button = ttk.Button(master=self.__action_frame, image=self.__play_icon,
command=self.on_enter_pressed, style='TButton', width=4)
self.__go_button_tooltip = Tooltip(self.__go_button, 'Execute corners detection')
self.__go_button.pack(side=tk.RIGHT)
# Get a second line in the actions area, so the checkbox will be under the entry
helper_frame = ctl.create_frame(master, tk.X)
self.__canny_checkbutton, self.__canny_check_var = ctl.create_checkbutton(helper_frame, 'Go through Canny Edge Detection', tk.LEFT)
self.__gauss_checkbutton, self.__gauss_check_var = ctl.create_checkbutton(helper_frame, 'Apply Gaussian Blur', tk.LEFT)
self.__gauss_checkbutton.select()
self.__popout_icon = tk.PhotoImage(file=os.path.abspath(os.path.join('resource', 'pop-out-icon.png')))
self.__popup_image_button = ttk.Button(master=helper_frame, image=self.__popout_icon, command=self.popup_image, style='TButton')
self.__popup_image_button_tooltip = Tooltip(self.__popup_image_button, 'Pops out the image with marks as a pyplot dialog')
self.__popup_image_button.pack(side=tk.RIGHT)
self.__popup_image_button['state'] = 'disabled'
def create_status_section(self, master):
"""
Status area contains a progress bar which we put at the bottom of the frame, to show a progress indication when
the algorithm is executing. (Can take some seconds)
We also add a label into the progress bar to use it as a status bar as well
:param master: Master dialog to add the frame to
:return: None
"""
self.__status_frame = ctl.create_frame(master, tk.X, 0, 0)
self.__status_frame.pack(fill=tk.X, side=tk.BOTTOM, anchor=tk.S)
# Add label into the layout
self.__style.layout('text.Horizontal.TProgressbar',
[('Horizontal.Progressbar.trough',
{'children': [('Horizontal.Progressbar.pbar', {'side': 'left', 'sticky': 's'})], 'sticky': 'swe'}),
('Horizontal.Progressbar.label', {'sticky': 'we'})])
self.__style.configure('text.Horizontal.TProgressbar', font=ctl.FONT_REGULAR)
self.__progress_bar = tk.ttk.Progressbar(master=self.__status_frame, orient=tk.HORIZONTAL, style='text.Horizontal.TProgressbar',
length=101, mode='determinate', value=0, maximum=101)
self.__progress_bar.pack(fill=tk.X)
self.update_status('Open an image using the magnifying button and click Play')
def create_workarea_section(self, master):
"""
Workarea contains a frame onto which we are displaying the plots when we finish executing Harris Detector algorithm
:param master: Master dialog to add the frame to
:return: None
"""
self.__figure_frame = ctl.create_frame(master, tk.BOTH)
self.__figure_frame.configure(background='white') # Make it white because I could not modify the PyPlot background
def update_status(self, text):
"""
Updates the text of the progress bar (status bar) with the specified text
:param text: The text to set into the status bar
:return: None
"""
self.__style.configure('text.Horizontal.TProgressbar', text=' ' + text)
self.progress_text_format = ' ' + text + ' {0}%'
print(text)
def update_progress(self, progress):
"""
Updates the progress bar with a progress, so user can see how much time remains
:param progress: The progress sto set to the progressbar
:return: None
"""
progress = np.min([100, int(progress)])
self.__progress_bar['value'] = progress
self.__style.configure('text.Horizontal.TProgressbar', text=self.progress_text_format.format(progress))
def start_progress(self):
"""
Prepare for starting a progress in the progress bar, for visualizing process.
We also start checking the job queue to detect if the algorithm has finished its execution so we can display the outcome
:return: None
"""
self.set_user_components_state('disabled')
self.__image = self.__processed_image = None
self.__running = True
self.periodically_check_outcome()
if self.__figure is not None:
self.__figure.set_visible(False)
self.__figure.clear()
self.__figure = None
# Remove all children so we will not have issues with pyplot's painting
for child in self.__figure_frame.winfo_children():
child.destroy()
def stop_progress(self):
"""
Stops the progress of the progress bar, and the periodic check of the job queue
:return: None
"""
# self.__progress_bar.stop()
self.__progress_bar['value'] = 0
self.__running = False
self.set_user_components_state('normal')
def set_user_components_state(self, new_state):
"""
Sets the state of user components to 'normal' or 'disabled'.
When the algorithm is running in background, we disable the user components so there won't be a mess
:param new_state: The new state to set. Can be one of 'normal' or 'disabled'
:return: None
"""
self.__go_button['state'] = new_state
self.__open_file_button['state'] = new_state
self.__file_path_entry['state'] = new_state
self.__canny_checkbutton['state'] = new_state
self.__gauss_checkbutton['state'] = new_state
self.__popup_image_button['state'] = new_state
def open_file_action(self):
"""
Displaying a file open dialog in image selecting mode, to select an image for executing the algorithm on
:return: None
"""
file_name = tkfiledialog.askopenfilename(filetypes=[("Image File", '.jpg .jpeg .png .bmp'), ("Video File", '.mp4 .avi .wmv')])
self.__file_path_entry.delete(0, tk.END)
self.__file_path_entry.insert(0, file_name)
def show_settings_dialog(self):
"""
Displaying Settings dialog so user can customize the algorithm settings
:return: None
"""
settings = SettingsDialog(self.master, self.settings).result
if settings:
self.settings = settings
self.settings.save()
def on_enter_pressed(self, event=None):
"""
This event is raised whenever user presses the Go button or Enter
In this case we are going to execute Harris Detector algorithm using the selected image, in case there is a valid selection.
:param event: The keypress event (We register on the master frame)
:return: None
"""
file_path = self.__file_path_entry.get().strip()
if file_path == '':
messagebox.showerror('Illegal Input', 'Please select a file first')
return None
if not os.path.exists(file_path) or not os.path.isfile(file_path):
messagebox.showerror('Illegal Input', 'Selected file is not a file or it does not exist:\n{}'.format(file_path))
return None
if not self.__running:
if is_video(file_path):
self.play_selected_video(file_path)
else:
self.start_progress()
# Run it in background so the progress bar will not get blocked. (We cannot block te gui thread)
t = Thread(target=self.execute_harris_detector, args=(file_path,))
t.start()
else:
print('Already running. Cannot run multiple detections in parallel.')
def execute_harris_detector(self, path):
"""
The job of executing Harris Detector algorithm.
It takes time so we have a specific action for that, such that we can run it using a background thread
:param path: Path to the selected image
:return: None
"""
self.__image, self.__processed_image = corners_and_line_intersection_detector(cv2.imread(path),
lambda text: self.update_status(text),
lambda progress: self.update_progress(progress),
bool(self.__canny_check_var.get()),
bool(self.__gauss_check_var.get()),
self.settings)
if self.__image is None or self.__processed_image is None:
self.__error = True
def show_images(self, img, processed_image):
"""
When Harris Detector job has finished we display the results as embedded figure
:param img: Original image
:param processed_image: Image with corner marks
:return: None
"""
self.__figure = Figure(figsize=(5, 5), dpi=100)
axes = self.__figure.add_subplot(121)
axes.imshow(np.uint8(img[:, :, ::-1]))
axes = self.__figure.add_subplot(122)
axes.imshow(np.uint8(processed_image[:, :, ::-1]))
self.__figure.subplots_adjust(0.05, 0.2, 0.95, 0.9, 0.2, 0.2)
canvas = FigureCanvasTkAgg(self.__figure, self.__figure_frame)
canvas.draw()
canvas.get_tk_widget().pack(side=tk.BOTTOM, fill=tk.BOTH, expand=True)
toolbar = NavigationToolbar2Tk(canvas, self.__figure_frame)
toolbar.update()
canvas._tkcanvas.pack(side=tk.TOP, fill=tk.BOTH, expand=True)
def periodically_check_outcome(self):
"""
Check every 200 ms if there is something new to show, which means a harris_detector worker has finished
and we can pick the output and show it in the GUI
"""
if self.__error:
self.__error = False
self.stop_progress()
messagebox.showerror('Error', 'Error has occurred while trying to detect corners')
self.__image = self.__processed_image = None
return None
if self.__image is not None and self.__processed_image is not None:
self.stop_progress()
# Plot the images, embedded within our dialog rather than popping up another dialog.
self.show_images(self.__image, self.__processed_image)
if self.__running:
self.master.after(50, self.periodically_check_outcome)
def popup_image(self):
"""
Action corresponding to when user presses the popup button, to plot the outcome outside the main window
:return: None
"""
if self.__processed_image is not None:
plt.figure('Corners Detector')
plt.tight_layout(pad=0.5)
plt.imshow(np.uint8(self.__processed_image[:, :, ::-1]))
plt.subplots_adjust(0.05, 0.05, 0.95, 0.95, 0.1, 0.1)
figure_manager = plt.get_current_fig_manager()
figure_manager.window.iconbitmap(os.path.abspath(os.path.join('resource', 'corners-icon.ico')))
try:
# In case there would be an issue on Ubuntu or something like that, just catch the error and use resize
figure_manager.window.state('zoomed')
except Exception as e:
print('Error has occurred while trying to show window as maximized. Fallback to resize strategy:', str(e))
figure_manager.resize(*figure_manager.window.maxsize())
figure_manager.window.wm_geometry("+0+0")
plt.show()
def play_selected_video(self, file_path):
"""
This method will load a video from file and display it in cv2's window.
We will execute corners detection, using the application settings, over the frames of the selected video and draw them live.
:param file_path: Path to the video file to load
:return: None
"""
font = cv2.FONT_HERSHEY_COMPLEX
bottom_left_corner_of_text = (10, 30)
font_scale = 1
font_color = (0, 255, 0)
lineType = 1
cap = cv2.VideoCapture(file_path)
while cap.isOpened():
ret, frame = cap.read()
if frame is None:
break
# Don't send listeners while playing a video in order to avoid of flooding the console
image, processed_image = corners_and_line_intersection_detector(frame,
is_using_canny=bool(self.__canny_check_var.get()),
is_applying_gauss=bool(self.__gauss_check_var.get()),
settings=self.settings)
self.__progress_bar['value'] = 0
if processed_image is not None:
cv2.putText(processed_image, 'Press q to stop', bottom_left_corner_of_text, font, font_scale, font_color, lineType)
cv2.imshow('Corners Detector', processed_image)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
cap.release()
