def extrapolate_metadata(directory_path, database_name, mongo_container):
metadata = load_exif_data(directory_path)
# Kalman
kalman = KalmanFilter()
initial_image_metadata = metadata[0]
initial_utm = utm.from_latlon(
initial_image_metadata.get("latitude", None),
initial_image_metadata.get("longitude", None))
init_data = np.array([initial_utm[0], initial_utm[1], 0, 0]).T
init_noise_vec = np.array([
initial_image_metadata.get("accuracy", None),
initial_image_metadata.get("accuracy", None), 2, 2
]).T
kalman.initialize(init_data=init_data,
init_noise_vec=init_noise_vec,
start_timestamp=initial_image_metadata.get("time", None))
updated_data = update_using_kalman(kalman, metadata)
geographic_db = GeographicMongoDB(database_name, mongo_container)
insert_data_to_mongo(geographic_db, updated_data)
geographic_db = GeographicMongoDB(database_name, container_name)
with open(gpx_file_address) as gpx_file:
gpx_data = gpxpy.parse(gpx_file)
video_kalman = KalmanFilter(location_noise=4, speed_noise=3)
way_points = gpx_data.waypoints
utm_data = utm.from_latlon(way_points[0].latitude, way_points[0].longitude)
print(utm_data)
video = cv.VideoCapture(video_address)
fake_timestamp = way_points[0].time.timestamp()
fake_time_start = fake_timestamp
start_real_time = time.time()
prev_time = start_real_time
video_kalman.initialize(
np.array([utm_data[0], utm_data[1], 0, 0]).T,
np.array([4, 4, 3, 3]).T, start_real_time)
index = 1
video.set(1, 100)
current_image = ""
while index < len(way_points):
if fake_timestamp >= way_points[index].time.timestamp():
if way_points[index].latitude > 0:
utm_data = utm.from_latlon(way_points[index].latitude,
way_points[index].longitude)
video_kalman.update_estimations(np.array(utm_data[:2]).T)
print(way_points[index])
index += 1
success, image = video.read()
if not success: