Merge images¶

This function takes a set of images with a known overlap and concatenates them in series to create a final single image. Image paths are provided in a list (e.g. read_dataset) and must be named in the order they should be merged (ie. prefix_1.png, prefix_2.png, etc.). There are 4 methods for handling the regions that overlap.

plantcv.transform.merge_images(paths_to_imgs, overlap_percentage, direction = "vertical", method = "stacked")

returns combined_image

• Parameters:

  • paths_to_imgs - a list of image paths
  • overlap_percentage - amount of overlap between successive images (see example below for how to calculate this value)
  • direction - vertical (default), or horizontal, determines in which direction successive images should be added
  • method - Available options are stacked, random, average, and gradual and dictate how the overlap region should be handled
    • 'stacked' : (default), image i+1 stacks on top of image i
    • 'random' : randomly choose either image i or i+1 to go on top
    • 'average' : pixels are averaged between image i values and image i+1 values
    • 'gradual' : pixels are averaged with a weight that corresponds to proximity to image i or image i+1

• Context:

  • Used to concatenate images in a series or scan, such as from minirhizotrons
  • Output image can be the start of any workflow

• Example use:

Vertical image with gradual method example

Input images

from plantcv import plantcv as pcv

# Set global debug behavior to None (default), "print" (to file), or "plot" (Jupyter Notebooks or X11)
pcv.params.debug = "plot"

# Merge images vertially with a 30% overlap by averaging the overlap pixels
images = pcv.io.read_dataset(source_path="./input_images/")
merged_img = pcv.transform.merge_images(paths_to_imgs=images, overlap_percentage=30, direction="vertical", method="gradual")

To calculate the overlap percentage, use the following formula:

# Calculate the overlap percentage
# overlap = amount of overlap in measured units (e.g. inches)
# dpu = dots per unit of the image (e.g. 100), units must match overlap units
# width or height = width or height of the image in pixels depending on orientation
#
# percent = ((overlap * dpu) / width or height) * 100
overlap = ((0.5 * 100) / img.shape[0]) * 100

Merged image

Source Code: Here