File size: 2,509 Bytes
94184e8 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 |
import torch
import torchvision.transforms as transforms
import pandas as pd
from PIL import Image
import os
import numpy as np
def prepare_data(csv_path: str):
"""
Load images and GPS coordinates from CSV.
Args:
csv_path: Path to CSV file with columns: image_path/filepath/image/path/file_name,
Latitude/latitude/lat, Longitude/longitude/lon
Returns:
X: torch.Tensor of shape (N, 3, 224, 224) - normalized images
y: torch.Tensor of shape (N, 2) - raw lat/lon in degrees
"""
# Read CSV
df = pd.read_csv(csv_path)
# Find the correct column names (case-insensitive)
image_col = None
lat_col = None
lon_col = None
for col in df.columns:
col_lower = col.lower()
if col_lower in ['image_path', 'filepath', 'image', 'path', 'file_name']:
image_col = col
elif col_lower in ['latitude', 'lat']:
lat_col = col
elif col_lower in ['longitude', 'lon']:
lon_col = col
if image_col is None or lat_col is None or lon_col is None:
raise ValueError(f"Could not find image, latitude, or longitude columns in CSV")
# Define transform (same as used during training)
transform = transforms.Compose([
transforms.Resize((224, 224)),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406],
std=[0.229, 0.224, 0.225])
])
# Load images and GPS coordinates
images = []
gps_coords = []
csv_dir = os.path.dirname(csv_path)
for idx, row in df.iterrows():
# Get image path
img_path = row[image_col]
# Handle relative paths
if not os.path.isabs(img_path):
img_path = os.path.join(csv_dir, img_path)
try:
# Load and transform image
image = Image.open(img_path).convert('RGB')
image = transform(image)
images.append(image)
# Get GPS coordinates (raw, in degrees)
lat = float(row[lat_col])
lon = float(row[lon_col])
gps_coords.append([lat, lon])
except Exception as e:
print(f"Warning: Could not load image {img_path}: {e}")
continue
# Convert to tensors
X = torch.stack(images) # Shape: (N, 3, 224, 224)
y = torch.tensor(gps_coords, dtype=torch.float32) # Shape: (N, 2)
return X, y
|