How to perform PCA and UMAP dimensionality reduction with multiple images together

Question

I have implemented PCA and UMAP dimensionality reduction for a single hyperspectral image. I didn't any problem. But now I have multiple hyperspectral images(more than 200). I would perform PCA and UMAP dimensionality reduction on datasets that contain hyperspectral multiple images. Please find the dataset shape look like - Firstly I have extracted the data frame from a single image -

import pandas as pd
import numpy as np

def extract_pixels(datacube, labelmap):
  q = datacube.reshape(-1, datacube.shape[2])
  df = pd.DataFrame(data = q)
  df = pd.concat([df, pd.DataFrame(data = labelmap.ravel())], axis=1)
  df.columns= [f'band{i}' for i in range(1, 1+datacube.shape[2])]+['class']
  df.to_csv('Dataset.csv')
  return df
  
df = extract_pixels(datacube, labelmap)

Then I apply PCA to algorithm to above df

from sklearn.decomposition import PCA

pca = PCA(n_components = 15)

principalComponents = pca.fit_transform(df.iloc[:, :-1].values)

My question is how do I apply PCA and UMAP to all images together. Practically It is not efficient to apply PCA and UMAP to all 200 images.