如何在图像之间传输像素颜色和位置

Question

I am trying to take the points in this image:

data.png

and transfer them to this US map outline:

us_outline.png

but I am struggling with it.

I am trying to use a method in which I read the color and coordinate location of the non-green pixels from 'data.png', store them in a list, and them place those pixels onto 'us_outline.png' based of off their location.

Here is the code I have so far:

#IMPORTS
from __future__ import division
import math
import numpy as np
from PIL import Image
import matplotlib.pyplot as mplot

#List of pixels from data.png
pixels = []

height = 140
width = 200

#Read in data from data.png
data = Image.open( "data.png" )
data = data.convert('RGB')

for row in range(0,height): #loops over the number of rows in the image
    for col in range(0,width): # loops over the number of columns in the current row
        r,g,b = data.getpixel((row,col))
        rgb = []
        rgb.append(r)
        rgb.append(g)
        rgb.append(b)
        if rgb != [0,255,0]:
            pixels.append(rgb)

But doing so results in an error: IndexError: image index out of range

I have also tried this:

#Convert to float32 format
data_image = np.float32(data)

#Reads in data points from data.png and appends them to a list
for row in range(len(data_image)): #loops over the number of rows in the image
    for col in range(len(data_image[row])): # loops over the number of columns in the current row
        pixel = data_image[row][col] #Assigns pixel at row and column to a variable
        if pixel != [0,255,0,255]: #If pixel is not green (a.k.a it is a data point)
            pixels.append(pixel) #Set key to the location of pixel and set value to pixel color

#Read in data from us_outline.png     
img2 = Image.open( "us_outline.png" )
usmap = img2.load()
#Convert to float32 format
usmap_image = np.float32(usmap)

#Writes data from pixels list to US map
for row in range(len(usmap_image)): #loops over the number of rows in the image
    for col in range(len(usmap_image[row])): # loops over the number of columns in the current row 
        for pixel in range(len(pixels)):
            if pixels[row][col] == usmap_image[row][col]:
                usmap_image[row][col] = pixels[row][col]

usmap_image = np.uint8( usmap_image ) 

but doing so results in errors in lines 21 and 22

I have also tried simply adding the two images together, but that yielded a weird result.

I have tried many methods and I can't figure out how to get it to work. Please help!

Thanks in advance

Answer 1

In your first piece of code, you just need to swap row and col to read in the pixels correctly. Line 18 becomes

r,g,b = data.getpixel((col, row))

Otherwise, the following piece of code accomplishes your goal and is a bit more succinct:

import numpy as np
import matplotlib.pyplot as plt

# find indices of non-green pixels
data = plt.imread('data.png')
green = np.zeros_like(data)
green[:,:,1] = 1. # plt.imread for some bizarre reason returns rgb values between 0.-1. for the given pngs, not 0-255!
x, y = np.where(np.any(data != green, axis=-1))

# plot non-green pixels on us outline
us = plt.imread('us_outline.png')
us[x,y] = data[x,y]

plt.imshow(us)
plt.show()