A script to allow for classification of snow interception in a forested area.

classify_snow.py

#!/usr/bin/env python
"""
Script to loop through images and categorize them in some way.
"""

import os
import tkinter
import pandas as pd
from PIL import Image
from PIL import ImageTk

# Maximum dimensions of image
MAX_WIDTH = 800
MAX_HEIGHT = 600

# Where to get the images from
DIRNAME = './Original_Images'

# Store the data
EVENTS = []
DATES = []

# Map the data
MAP = {1: 'open',
       2: 'light snow',
       3: 'heavy snow'}


def button_click_exit_mainloop(event):
    EVENTS.append(int(event.char))
    event.widget.quit()


def main():
    root = tkinter.Tk()
    root.bind("<Key>", button_click_exit_mainloop)
    root.geometry('+{}+{}'.format(100, 100))
    dirlist = os.listdir(DIRNAME)
    old_label_image = None
    for f in dirlist:
        try:
            image = Image.open('{}/{}'.format(DIRNAME, f))
            img_date = image._getexif()[36867]
            DATES.append(img_date)
            width = min([MAX_WIDTH, image.size[0]])
            height = min([MAX_HEIGHT, image.size[1]])
            image = image.resize((width, height), Image.ANTIALIAS)
            root.geometry('{}x{}'.format(width, height))
            tkpi = ImageTk.PhotoImage(image)
            label_image = tkinter.Label(root, image=tkpi)
            label_image.place(x=0, y=0, width=width, height=height)
            root.title(f)
            if old_label_image is not None:
                old_label_image.destroy()
            old_label_image = label_image
            root.mainloop()
        except:
            # Skip things that aren't images
            pass
    # Write out the data
    df = pd.DataFrame({'raw': EVENTS,
                       'snow_level': [MAP[i] for i in EVENTS],
                       'date': DATES})
    df.to_csv('./processed.csv')
    # Make sure that we kill the interface when we're done
    root.destroy()


if __name__ == '__main__':
    main()