Automatically crop scanned images

Automatically cropping scanned images involves detecting and removing unwanted borders or background from the edges of a scanned document or photo. This can be accomplished through various methods, depending on the tools or programming languages used. Below are different approaches to achieve this:

---

1. Using Python with OpenCV

Requirements: opencv-python, numpy

bash
CopyEdit
pip install opencv-python numpy

Code Example:

python
CopyEdit
import cv2
import numpy as np

def auto_crop_image(image_path, output_path):
    image = cv2.imread(image_path)
    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)

    _, thresh = cv2.threshold(gray, 240, 255, cv2.THRESH_BINARY_INV)

    contours, _ = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

    if not contours:
        print("No contours found!")
        return

    # Find bounding box for all contours
    x, y, w, h = cv2.boundingRect(np.vstack(contours))
    cropped = image[y:y+h, x:x+w]

    cv2.imwrite(output_path, cropped)

# Example usage
auto_crop_image('scanned_input.jpg', 'cropped_output.jpg')

---

2. Using ImageMagick (Command Line Tool)

Install:

• Linux: sudo apt install imagemagick

• macOS: brew install imagemagick

• Windows: Install via official site

Command:

bash
CopyEdit
convert scanned_input.jpg -fuzz 10% -trim +repage cropped_output.jpg

• -fuzz 10% allows for color variation tolerance.

• -trim removes borders of the same or similar color.

• +repage resets image canvas.

---

3. Using Adobe Acrobat Pro (Manual GUI)

• Open your scanned PDF.

• Select Edit PDF tool.

• Use Crop Pages option under the Set Page Boxes.

• Adjust manually or use auto-detect settings.

---

4. Using Online Tools

Some popular options:

• iLovePDF

• ScanWritr

• Online2PDF

These allow basic cropping with some automation, though advanced batch processing may require a paid account.

---

5. Using Tesseract + OpenCV (For OCR & Layout Detection)

You can combine Tesseract for layout recognition and OpenCV for cropping around text regions:

python
CopyEdit
import cv2
import pytesseract

image = cv2.imread('scanned_input.jpg')
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
custom_config = r'--oem 3 --psm 6'
boxes = pytesseract.image_to_boxes(gray, config=custom_config)

height, _, _ = image.shape
x_min, y_min, x_max, y_max = float('inf'), float('inf'), 0, 0

for b in boxes.splitlines():
    b = b.split(' ')
    x, y, w, h = int(b[1]), int(b[2]), int(b[3]), int(b[4])
    x_min = min(x_min, x)
    y_min = min(y_min, height - h)
    x_max = max(x_max, w)
    y_max = max(y_max, height - y)

cropped = image[y_min:y_max, x_min:x_max]
cv2.imwrite('ocr_cropped_output.jpg', cropped)

---

Best Practices for Batch Cropping

• For multiple images, loop over a directory.

• Normalize DPI and contrast before cropping for better accuracy.

• Combine deskewing and denoising if images are skewed.

---

When to Use What

Let me know if you want a GUI version or need integration with another language or framework.