Crop and merge images using Python (OpenCV)

Manually cropping and merging a large number of images can be a boring and time-consuming task. We can automate this boring stuff with Python and the OpenCV library¹. In this tutorial, we’ll explore how to crop and merge multiple images using Python and OpenCV.

Scenario

Suppose we have a collection of similar images, and we want to crop them at a certain location before merging every two images together. The images are 600px X 600px in size, and there’s a fixed region where we want to perform the cropping.

For example, let’s consider the following two brain image figures:

First image: this will be placed on left side

Second image: this will be placed on right side

Our goal is to crop the outer white spaces and the text on top, and then merge the two images into a single image like the following:

Manually merged image: this image is manually merged from the above two images

Automate the cropping and merging of images using Python

Step 1: Determine the coordinates in the images for cropping

To crop an image using code, we need to know the exact pixel location and size of the region to crop. In this case, we’ll use an online tool to get the coordinates where we need to crop ².

The top-left corner of the image is represented by the coordinate (0, 0). In our example, the desired position to start cropping is (130, 140), with a size of 350px X 400px.

Calculating the start coordinates and desired image size using an online crop tool

Step 2: Use Python to crop and merge the images

• First, we need to install the opencv-python package, which will be used to crop and merge the images:

  pip install opencv-python
  

• The following code will read the images, crop the specified regions, merge them side by side, and save the result as a new image file.

  import cv2
  
  
  def crop_and_merge():
  combined_image_file = "paper_exp/survey_1/1.png"
  left_image_file = "paper_exp/survey_1/1_r_1_645_mid.png"
  right_image_file = "paper_exp/survey_1/1_r_1_2500_mid.png"
  
  # Load the images
  left_image = cv2.imread(left_image_file)
  right_image = cv2.imread(right_image_file)
  
  # Set the crop parameters
  x_start = 130
  y_start = 140
  crop_width = 350
  crop_height = 400
      
  # Crop the images
  cropped_left_image = left_image[y_start:y_start + crop_height, x_start:x_start + crop_width]
  cropped_right_image = right_image[y_start:y_start + crop_height, x_start:x_start + crop_width]
      
  # Combine the images horizontally (side by side)
  combined_image = cv2.hconcat([cropped_left_image, cropped_right_image])
      
  # Save the combined image
  cv2.imwrite(combined_image_file, combined_image)
  print(f"Saved: {combined_image_file}")
  
  # Call the function to crop and merge the images    
  crop_and_merge()    
  

In this code:

• We first load the left and right images using cv2.imread().
• We set the crop parameters (x_start, y_start, crop_width, and crop_height) based on the coordinates obtained in Step 1.
• We crop the images using array slicing.
• We combine the cropped images horizontally using cv2.hconcat().
• Finally, we save the combined image using cv2.imwrite().

Automatically cropped and merged image: This image is automatically created from the above two images

By automating the process of cropping and merging images with Python and OpenCV, we can save a significant amount of time and effort, especially when dealing with a large number of images.

References: