Datasets:
Tasks:
Text Classification
Formats:
text
Languages:
Karo (Brazil)
Size:
10K - 100K
Tags:
driving
License:
import time | |
import numpy as np | |
from PIL import ImageGrab | |
import keyboard | |
def capture_screen(): | |
start_time = 0 | |
filename = '' | |
while True: | |
if keyboard.is_pressed('F7') and start_time == 0: # Check if F7 is pressed to start capturing | |
start_time = int(time.time()) # Record start time | |
filename = 'trainingData_{}.txt'.format(start_time) | |
break | |
while True: | |
try: | |
if keyboard.is_pressed('F8'): # Check if F7 is pressed to stop capturing | |
break | |
# Capture entire screen | |
screenshot = ImageGrab.grab() | |
# Resize to 128x128 | |
screenshot = screenshot.resize((128, 64)) | |
# Convert to grayscale | |
grayscale_img = screenshot.convert('L') | |
# Convert image to numpy array | |
grayscale_array = np.array(grayscale_img) | |
# Normalize pixel values between 0 and 1 and round to 2 decimal places | |
grayscale_array = np.round(grayscale_array / 255.0, 1) | |
# Flatten array to 1D | |
flattened_array = grayscale_array.flatten() | |
# Save the image with two decimal place precision | |
screenshot.save('captured_image.png') | |
# Capture WASD keys | |
keys = [ | |
int(keyboard.is_pressed(key)) for key in ['w', 'a', 's', 'd'] | |
] | |
# Append to training data file with Unix time | |
with open(filename, 'a') as f: | |
f.write("{data}\n") | |
f.write(str(flattened_array.tolist()) + '\n') | |
f.write("{action}\n") | |
f.write(str(keys) + '\n') | |
# Wait for 0.1 seconds | |
time.sleep(0.1) | |
except KeyboardInterrupt: | |
print("Stopping screen capture.") | |
break | |
if __name__ == "__main__": | |
while True: | |
capture_screen() | |