Watermeter / extract.py
skywalker290's picture
Upload folder using huggingface_hub
730a0fa verified
import os
import cv2
import numpy as np
from tqdm import tqdm
def extract_masked_region(img: np.ndarray,
msk: np.ndarray,
threshold: int = 127,
crop: bool = True) -> np.ndarray:
# Binarize mask
_, bin_mask = cv2.threshold(msk, threshold, 255, cv2.THRESH_BINARY)
# Apply mask
result = cv2.bitwise_and(img, img, mask=bin_mask)
if crop:
ys, xs = np.where(bin_mask == 255)
if ys.size and xs.size:
y1, y2 = ys.min(), ys.max()
x1, x2 = xs.min(), xs.max()
result = result[y1:y2+1, x1:x2+1]
return result
def batch_process(collage_dir: str,
mask_dir: str,
out_dir: str):
os.makedirs(out_dir, exist_ok=True)
files = [f for f in os.listdir(collage_dir)
if f.lower().endswith(('.png','.jpg','.jpeg','.bmp','tif','tiff'))]
for fname in tqdm(files, desc="Processing images"):
img_path = os.path.join(collage_dir, fname)
mask_path = os.path.join(mask_dir, fname)
out_path = os.path.join(out_dir, fname)
if not os.path.isfile(mask_path):
tqdm.write(f"⚠️ mask not found for {fname}, skipping")
continue
img = cv2.imread(img_path, cv2.IMREAD_COLOR)
msk = cv2.imread(mask_path, cv2.IMREAD_GRAYSCALE)
if img is None or msk is None:
tqdm.write(f"⚠️ failed to load {fname} or its mask, skipping")
continue
cropped = extract_masked_region(img, msk, threshold=127, crop=True)
cv2.imwrite(out_path, cropped)
tqdm.write("✅ All done!")
if __name__ == "__main__":
# adjust these paths as needed:
collage_folder = "/home/darth/#/WaterMeters/images"
mask_folder = "/home/darth/#/WaterMeters/masks"
output_folder = "/home/darth/#/WaterMeters/cropped"
batch_process(collage_folder, mask_folder, output_folder)
# import cv2
# import numpy as np
# def extract_masked_region(image_path: str,
# mask_path: str,
# threshold: int = 127,
# crop: bool = True):
# """
# Loads an image and its mask, applies the mask, and returns the resulting image.
# If crop=True, it also crops to the bounding box of the white region in the mask.
# :param image_path: Path to the original BGR image.
# :param mask_path: Path to the grayscale mask (white=keep, black=discard).
# :param threshold: Grayscale threshold to binarize the mask (0–255).
# :param crop: If True, crop to the mask's bounding box.
# :return: A BGR image with only the masked region (and optionally cropped).
# """
# # 1. Load images
# img = cv2.imread(image_path, cv2.IMREAD_COLOR)
# msk = cv2.imread(mask_path, cv2.IMREAD_GRAYSCALE)
# if img is None or msk is None:
# raise FileNotFoundError("Could not load image or mask. Check your paths.")
# # 2. Binarize mask
# _, bin_mask = cv2.threshold(msk, threshold, 255, cv2.THRESH_BINARY)
# # 3. Apply mask
# result = cv2.bitwise_and(img, img, mask=bin_mask)
# if crop:
# # 4. Find bounding box of white region
# ys, xs = np.where(bin_mask == 255)
# if ys.size and xs.size:
# y1, y2 = ys.min(), ys.max()
# x1, x2 = xs.min(), xs.max()
# result = result[y1:y2+1, x1:x2+1]
# return result
# if __name__ == "__main__":
# out = extract_masked_region("/home/darth/#/WaterMeters/collage/id_1_value_13_116.jpg", "/home/darth/#/WaterMeters/masks/id_1_value_13_116.jpg")
# cv2.imwrite("meter_extracted.png", out)
# print("Saved extracted region to meter_extracted.png")