Using the PIL (Python Imaging Library) you can easily load images and process them. Using the grayscale conversion, you can convert RGB to grayscale, which should be easier to detect levels. If you want to threshold the image (to detect the white boards), there is a point() function which lets you map the colors.
On the other hand, you could write a simple program, which lets you
- select, and show an image
- mark the area where the boards are
- crop the image
- apply tesseract or whatever,
- save the image with the detected number
That should facilitate the process a lot! Writing this should be relatively easy using TkInter, PyGTK, PyQt or some other windowing toolkit.
EDIT: I was needing a similar program to categorize images here - though not OCRing them. So I finally decided this was as good a time as any and made a first try (with OCR!). Make a backup of your images before trying it out!
Quick manual:
- Top left: Select the work folder, image list will appear if any images in the folder.
- Select image. Select area of image with number. Coordinates will appear in lower left corner, and program will call Tesseract.
- Edit - if necessary - the number OCRd in the dialog.
- Click Ok to accept - image will be renamed.
Here's the pre-alpha program:
#!/usr/bin/env python
# -*- coding: utf-8 -*-
#
# test_pil.py
#
# Copyright 2015 John Coppens
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
# MA 02110-1301, USA.
#
#
import pygtk
import gtk
import glob
import os.path as osp
from os import rename
import re
import subprocess as sp
temp_image = "/tmp/test_pil.png"
image_re = """\.(?:jpe?g|png|gif)$"""
class RecognizeDigits():
def __init__(self):
pass
def process(self, img, x0, y0, x1, y1):
""" Receive the gtk.Image, and the limits of the selected area (in
window coordinates!)
Call Tesseract on the area, and give the possibility to edit the
result.
Returns None if NO is pressed, and the OCR'd (and edited) text if OK
"""
pb = img.get_pixbuf().subpixbuf(x0, y0, x1-x0, y1-y0)
pb.save(temp_image, "png")
out = sp.check_output(("tesseract", temp_image, "stdout", "-psm 7", "digits"))
out = out.replace(" ", "").strip()
dlg = gtk.MessageDialog(type = gtk.MESSAGE_QUESTION,
flags = gtk.DIALOG_MODAL,
buttons = gtk.BUTTONS_YES_NO,
message_format = "The number read is:")
entry = gtk.Entry()
entry.set_text(out)
dlg.get_message_area().pack_start(entry)
entry.show()
response = dlg.run()
nr = entry.get_text()
dlg.destroy()
if response == gtk.RESPONSE_YES:
return nr
else:
return None
class FileSelector(gtk.VBox):
""" Provides a folder selector (at the top) and a list of files in the
selected folder. On selecting a file, the FileSelector calls the
function provided to the constructor (image_viewer)
"""
def __init__(self, image_viewer):
gtk.VBox.__init__(self)
self.image_viewer = image_viewer
fc = gtk.FileChooserButton('Select a folder')
fc.set_action(gtk.FILE_CHOOSER_ACTION_SELECT_FOLDER)
fc.connect("selection-changed", self.on_file_set)
self.pack_start(fc, expand = False, fill = True)
self.tstore = gtk.ListStore(str)
self.tview = gtk.TreeView(self.tstore)
self.tsel = self.tview.get_selection()
self.tsel.connect("changed", self.on_selection_changed)
renderer = gtk.CellRendererText()
col = gtk.TreeViewColumn(None, renderer, text = 0)
self.tview.append_column(col)
scrw = gtk.ScrolledWindow()
scrw.add(self.tview)
self.pack_start(scrw, expand = True, fill = True)
def on_file_set(self, fcb):
self.tstore.clear()
self.imgdir = fcb.get_filename()
for f in glob.glob(self.imgdir + "/*"):
if re.search(image_re, f):
self.tstore.append([osp.basename(f)])
def on_selection_changed(self, sel):
model, itr = sel.get_selected()
if itr != None:
base = model.get(itr, 0)
fname = self.imgdir + "/" + base[0]
self.image_viewer(fname)
class Status(gtk.Table):
""" Small status window which shows the coordinates for of the area
selected in the image
"""
def __init__(self):
gtk.Table.__init__(self)
self.attach(gtk.Label("X"), 1, 2, 0, 1, yoptions = gtk.FILL)
self.attach(gtk.Label("Y"), 2, 3, 0, 1, yoptions = gtk.FILL)
self.attach(gtk.Label("Top left:"), 0, 1, 1, 2, yoptions = gtk.FILL)
self.attach(gtk.Label("Bottom right:"), 0, 1, 2, 3, yoptions = gtk.FILL)
self.entries = {}
for coord in (("x0", 1, 2, 1, 2), ("y0", 2, 3, 1, 2),
("x1", 1, 2, 2, 3), ("y1", 2, 3, 2, 3)):
self.entries[coord[0]] = gtk.Entry()
self.entries[coord[0]].set_width_chars(6)
self.attach(self.entries[coord[0]],
coord[1], coord[2], coord[3], coord[4],
yoptions = gtk.FILL)
def set_top_left(self, x0, y0):
self.x0 = x0
self.y0 = y0
self.entries["x0"].set_text(str(x0))
self.entries["y0"].set_text(str(y0))
def set_bottom_right(self, x1, y1):
self.x1 = x1
self.y1 = y1
self.entries["x1"].set_text(str(x1))
self.entries["y1"].set_text(str(y1))
class ImageViewer(gtk.ScrolledWindow):
""" Provides a scrollwindow to move the image around. It also detects
button press and release events (left button), will call status
to update the coordinates, and will call task on button release
"""
def __init__(self, status, task = None):
gtk.ScrolledWindow.__init__(self)
self.task = task
self.status = status
self.drawing = False
self.prev_rect = None
self.viewport = gtk.Viewport()
self.viewport.connect("button-press-event", self.on_button_pressed)
self.viewport.connect("button-release-event", self.on_button_released)
self.viewport.set_events(gtk.gdk.BUTTON_PRESS_MASK | \
gtk.gdk.BUTTON_RELEASE_MASK)
self.img = gtk.Image()
self.viewport.add(self.img)
self.add(self.viewport)
def set_image(self, fname):
self.imagename = fname
self.img.set_from_file(fname)
def on_button_pressed(self, viewport, event):
if event.button == 1: # Left button: Select rectangle start
#self.x0, self.y0 = self.translate_coordinates(self.img, int(event.x), int(event.y))
self.x0, self.y0 = int(event.x), int(event.y)
self.status.set_top_left(self.x0, self.y0)
self.drawing = True
def on_button_released(self, viewport, event):
if event.button == 1: # Right button: Select rectangle end
#self.x1, self.y1 = self.translate_coordinates(self.img, int(event.x), int(event.y))
self.x1, self.y1 = int(event.x), int(event.y)
self.status.set_bottom_right(self.x1, self.y1)
if self.task != None:
res = self.task().process(self.img, self.x0, self.y0, self.x1, self.y1)
if res == None: return
newname = osp.split(self.imagename)[0] + '/' + res + ".jpeg"
rename(self.imagename, newname)
print "Renaming ", self.imagename, newname
class MainWindow(gtk.Window):
def __init__(self):
gtk.Window.__init__(self)
self.connect("delete-event", self.on_delete_event)
self.set_size_request(600, 300)
grid = gtk.Table()
# Image selector
files = FileSelector(self.update_image)
grid.attach(files, 0, 1, 0, 1,
yoptions = gtk.FILL | gtk.EXPAND, xoptions = gtk.FILL)
# Some status information
self.status = Status()
grid.attach(self.status, 0, 1, 1, 2,
yoptions = gtk.FILL, xoptions = gtk.FILL)
# The image viewer
self.viewer = ImageViewer(self.status, RecognizeDigits)
grid.attach(self.viewer, 1, 2, 0, 2)
self.add(grid)
self.show_all()
def update_image(self, fname):
self.viewer.set_image(fname)
def on_delete_event(self, wdg, data):
gtk.main_quit()
def run(self):
gtk.mainloop()
def main():
mw = MainWindow()
mw.run()
return 0
if __name__ == '__main__':
main()
![](\"https://i.stack.imgur.com/CO)
加载中...