Initial commit

Filter images with dry run and bounding box

filter_gps_images.py

@@ -0,0 +1,122 @@
+# Iterate over a folder with images
+# print the image gps coordinates
+# filter for images within a certain gps coordinates box
+# Use pillow to read image metadata
+import os
+import argparse
+import shutil
+from PIL import Image
+
+def parse_arguments():
+    parser = argparse.ArgumentParser(description="Filter images within a GPS coordinates box")
+    parser.add_argument("--input_folder", type=str, required=True, help="Path to the input folder containing images")
+    parser.add_argument("--output_folder", type=str, required=True, help="Path to the output folder to save filtered images")
+    gps_box_help = "Bounding box defined by min and max latitude and longitude" \
+                   " Swiss Alps region: --gps_box 46.0 8.0 48.0 9.0" \
+                   " Zurich area: --gps_box 47.2 8.3 47.6 8.7" \
+                   " Bernese Alps: --gps_box 46.4 7.8 46.8 8.3"
+    parser.add_argument("--gps_box", type=float, nargs=4, required=True, metavar=('LAT_MIN', 'LON_MIN', 'LAT_MAX', 'LON_MAX'), help=gps_box_help)
+    parser.add_argument("--dry_run", action="store_true", help="Enable dry run mode")
+    # Add help text for gps_box
+    parser.add_argument("--verbose", action="store_true", help="Enable verbose output")
+
+    return parser.parse_args()
+
+def get_gps_coordinates(image):
+    """Extract GPS coordinates from image EXIF data"""
+    exifdata = image.getexif()
+    
+    # Check if GPS info is available
+    if 34853 not in exifdata:
+        return None, None
+    
+    # Use get_ifd() to access the actual GPS data dictionary
+    try:
+        gps_info = exifdata.get_ifd(34853)
+    except (KeyError, TypeError):
+        return None, None
+    
+    # Check if gps_info is actually a dictionary with GPS data
+    if not isinstance(gps_info, dict) or len(gps_info) == 0:
+        return None, None
+    
+    def convert_to_degrees(value):
+        """Convert GPS coordinates to decimal degrees"""
+        try:
+            if isinstance(value, (list, tuple)) and len(value) >= 3:
+                d, m, s = value[0], value[1], value[2]
+                # Handle rational numbers (tuples like (40, 1) meaning 40/1)
+                if isinstance(d, tuple) and len(d) == 2:
+                    d = d[0] / d[1]
+                if isinstance(m, tuple) and len(m) == 2:
+                    m = m[0] / m[1]
+                if isinstance(s, tuple) and len(s) == 2:
+                    s = s[0] / s[1]
+                return float(d) + (float(m) / 60.0) + (float(s) / 3600.0)
+            return None
+        except (ValueError, TypeError, IndexError, ZeroDivisionError):
+            return None
+    
+    lat = None
+    lon = None
+    
+    try:
+        # Get latitude
+        if 2 in gps_info and 1 in gps_info:  # GPSLatitude and GPSLatitudeRef
+            lat = convert_to_degrees(gps_info[2])
+            if lat is not None and gps_info[1] == 'S':
+                lat = -lat
+        
+        # Get longitude  
+        if 4 in gps_info and 3 in gps_info:  # GPSLongitude and GPSLongitudeRef
+            lon = convert_to_degrees(gps_info[4])
+            if lon is not None and gps_info[3] == 'W':
+                lon = -lon
+    except (TypeError, KeyError):
+        # Handle cases where gps_info doesn't support 'in' operator or key access
+        return None, None
+            
+    return lat, lon
+
+def is_within_gps_box(image, gps_box):
+    # Get the image gps coordinates from the image metadata
+    latitude, longitude = get_gps_coordinates(image)
+    
+    if latitude is None or longitude is None:
+        print("No GPS data found in image")
+        return False
+        
+    print(f"Image GPS coordinates: {latitude}, {longitude}")
+    lat_min, lon_min, lat_max, lon_max = gps_box
+    return lat_min <= latitude <= lat_max and lon_min <= longitude <= lon_max
+
+def main():
+    args = parse_arguments()
+    
+    if not os.path.exists(args.output_folder):
+        os.makedirs(args.output_folder)
+    
+    for filename in os.listdir(args.input_folder):
+        if filename.lower().endswith(('.png', '.jpg', '.jpeg', '.bmp', '.tiff')):
+            image_path = os.path.join(args.input_folder, filename)
+            # Load the image metadata
+            image = Image.open(image_path)
+            
+            if image is None:
+                print(f"Failed to load image: {image_path}")
+                continue
+
+            if is_within_gps_box(image, args.gps_box):
+                if args.dry_run:
+                    print(f"Dry run: Image {filename} is within the GPS box")
+                    lat, lon = get_gps_coordinates(image)
+                    print(f"coordinates: {lat}, {lon}")
+                    continue
+                output_path = os.path.join(args.output_folder, filename)
+                shutil.copy2(image_path, output_path)
+                print(f"Saved filtered image: {output_path}")
+            else:
+                print(f"Image {filename} is outside the bounding box")
+
+if __name__ == "__main__":
+    main()