Sistema de Predicción de Biomasa de Cultivos

import torch
import cv2
import psycopg2
import pytorch_lightning as pl
from torch.utils.data import Dataset, DataLoader
import numpy as np

# Clase de preprocesamiento de imágenes
class BiomassImagePreprocessor(Dataset):
def __init__(self, image_data, drone_altitudes, reference_altitude=6, crop_size=256):
    self.image_data = image_data
    self.drone_altitudes = drone_altitudes
    self.reference_altitude = reference_altitude
    self.crop_size = crop_size

def preprocess_image(self, image, drone_altitude):
    scaling_factor = self.reference_altitude / drone_altitude
    crop_dim = int(self.crop_size * scaling_factor)
    height, width, _ = image.shape
    start_x = (width // 2) - (crop_dim // 2)
    start_y = (height // 2) - (crop_dim // 2)
    crop_img = image[start_y:start_y + crop_dim, start_x:start_x + crop_dim]
    upscale_img = cv2.resize(crop_img, (2048, 2048), interpolation=cv2.INTER_CUBIC)
    return upscale_img

def __len__(self):
    return len(self.image_data)

def __getitem__(self, idx):
    image = self.image_data[idx]
    drone_altitude = self.drone_altitudes[idx]
    preprocessed_image = self.preprocess_image(image, drone_altitude)
    # Convierte imagen a tensor
    preprocessed_image = torch.from_numpy(preprocessed_image).float().permute(2, 0, 1)
    return preprocessed_image

# Conexión a base de datos
def load_data_from_postgresql(query, conn_params):
conn = psycopg2.connect(**conn_params)
cursor = conn.cursor()
cursor.execute(query)
data = cursor.fetchall()
cursor.close()
conn.close()
return data

# Parámetros de conexión a PostgreSQL
conn_params = {
"dbname": "biomass_db",
"user": "postgres",
"password": "password",
"host": "localhost",
"port": "5432"
}

# Ejemplo de consulta a la base de datos
query = "SELECT image, drone_altitude FROM drone_images"