How to properly combine image using OpenCV

Question

I have satellite images like this:

• Images are all squares and have the size of 1024px.

• One images corresponds to 1600m of square land.

• Each adjacent image has overlaps of 1%.

So if I want to combine those images, I need to remove the overlapping area (grey) and leave the center area (green). The size of the remain are are 1584m (1600*99%) and 1013.76px.

The schematic showing the overlaps:

I don't need to consider advanced image stitching problems.

Suppose I have 10 satellite images, and I want to combine these horizontally.

Ideally, the combined image should be 15840m and 10137.6px. But if I need to chop the image first, so a single image become 1013px (should be 1013.76), the error is 0.76px. After combining 10 images like this, the combined image is 10130px, the error becomes 7.6px, which is a lot.

What can I do to get rid of this error?

The schematic of concatenating images:

Code would be somthing like this:

import cv2
import numpy as np

imgs = [np.zeros((1024, 1024, 3), dtype=np.uint8) for _ in range(10)] # satellite images

# (1024, 1024, 3)
print(imgs[0].shape)

for i in range(len(imgs)):
    # should be from 5.12 to 1018.88
    # but need to convert to interger
    p_start = int(1024 * 0.005)
    p_end = int(1024 - 1024 * 0.005)
    imgs[i] = imgs[i][p_start:p_end, p_start:p_end]

# (1013, 1013, 3)
print(imgs[0].shape)

img_conbine = cv2.hconcat(imgs)

# (1013, 10130, 3)
print(img_conbine.shape)