I have a folder with about 2000 satellite images of irrigation ponds.

I want to create a tensorflow model that, given a satellite image, will identify all the irrigation ponds that it contains.

import numpy as np
import os
import PIL
import PIL.Image
import tensorflow as tf
import pathlib

dataset_url = "E:/DeteccionObjetos/Muestra"
satellite= "E:/DeteccionObjetos/Orto.jpg"

data_dir = pathlib.Path(dataset_url )
image_count = len(list(data_dir.glob('*.png')))
roses = list(data_dir.glob('*')) 
batch_size = 32
img_height = 180
img_width = 180
train_ds = tf.keras.utils.image_dataset_from_directory(
  data_dir,
  validation_split=0.99,
  subset="training",
  seed=123,
  image_size=(img_height, img_width),
  batch_size=batch_size)

Create a tensorflow model that identifies irrigation ponds within an orthophoto