def reproject_and_save(dataset_dir, inference_dir, camera_name, seq):
seq_number = seq.split("/")[-1]
bbox_list = sorted(glob.glob(seq + "/*.txt"))
image_f_dir = os.path.join(dataset_dir, seq_number, "camera", camera_name)
intrinsics_f = read_json(image_f_dir + "/intrinsics.json")
intr_f = intrinscs_to_matrix(**intrinsics_f)
poses_f = read_json(image_f_dir + "/poses.json")
lidar_poses = read_json(
os.path.join(dataset_dir, seq_number, "lidar", "poses.json"))
output_dir = os.path.join(inference_dir, "image_bboxes", seq_number,
camera_name)
os.makedirs(output_dir, exist_ok=True)
for idx, bbox_path in enumerate(bbox_list):
name = os.path.splitext(os.path.basename(bbox_path))[0]
cam_T_lidar = get_extrinsic(lidar_poses[idx], poses_f[idx])
# lidar_file = os.path.join(
# dataset_dir, seq_number, "lidar_processed", name + ".bin"
# )
# lidar = load_lidar(lidar_file)
# lidar = lidar[:, :3]
# Load boxes
bboxes = load_bboxes(bbox_path, label_string=False)
labels = bboxes[:, -1]
lidar_corners_3d, _ = make_eight_points_boxes(bboxes[:, :-1])
orient_3d = get_orient(lidar_corners_3d)
lidar_corners_3d = project_boxes_to_cam(lidar_corners_3d, cam_T_lidar)
orient_3d = project_boxes_to_cam(orient_3d, cam_T_lidar)
# lidar = project_lidar_to_cam(lidar, cam_T_lidar)
# figure = visualize_lidar(lidar)
# figure = visualize_bboxes_3d(lidar_corners_3d, figure)
# project to image
image_path = os.path.join(image_f_dir, name + ".jpg")
image_f = np.asarray(Image.open(image_path)) / 255
box_2d = project_to_image(lidar_corners_3d, intr_f)
orient_2d = project_to_image(orient_3d, intr_f)
image = visualize_bboxes_on_image(image_f, box_2d, labels,
orient_2d) * 255
image_name = os.path.join(output_dir, name + ".png")
img = Image.fromarray(image.astype("uint8"))
img.save(image_name)
def report(
checkpoints_dir,
report_dir,
from_file=None,
full_report=False,
latex_pdf=True,
remove_plots=True,
):
"""
The generate json report and save it and also generate pdf report with plots
Arguments:
checkpoints_dir: ditectory containing all checkpoints from training
report_dir: directory to save reprot
from_file: if not None generate pdf report from specified .json file
full_report: generate .json file containing all information
from all epoch (not only from the best)
latex_pdf: if true generate pdf file otherwise only .json
remove_plots: if true removes all generated .png plots
"""
if from_file is None:
report_json = os.path.join(report_dir, "experiment_report.json")
graphics_dir = os.path.join(report_dir, "experiment_report")
os.makedirs(graphics_dir, exist_ok=True)
report_name = "experiment_report"
# Metrics to plot
plot_metrics = ["train_loss", "val_loss"]
main_metric = "val_loss" # Metric to define best epoch
report_dict = get_report_json(
checkpoints_dir,
report_dir,
plot_metrics,
main_metric,
full_report=full_report,
)
save_to_json(report_json, report_dict)
else:
report_dict = read_json(from_file)
graphics_dir = os.path.splitext(from_file)[0]
report_dir = "/".join(graphics_dir.split("/")[:-1])
report_name = graphics_dir.split("/")[-1]
os.makedirs(graphics_dir, exist_ok=True)
# Get metrics to plot
plot_dict = report_dict["plot_metrics"]
best_epoch = int(report_dict["best_epoch"])
# Create images of plots for: loss, miou, accuracy
plot_metric_to_file(graphics_dir,
plot_dict,
metric_name="loss",
best_epoch=best_epoch)
plot_learning_rate_to_file(graphics_dir, report_dict)
if latex_pdf:
generate_latex_pdf(graphics_dir, report_dir, report_dict, report_name)
if remove_plots:
shutil.rmtree(graphics_dir)
def preprocess_data(dataset_dir):
"""
The function visualizes data from pandaset.
Arguments:
dataset_dir: directory with Pandaset data
"""
shift_lidar = [25, 50, 2.5]
# Get list of data samples
search_string = os.path.join(dataset_dir, "*")
seq_list = sorted(glob.glob(search_string))
for seq in tqdm(seq_list, desc="Process sequences", total=len(seq_list)):
search_string = os.path.join(seq, "lidar", "*.pkl.gz")
lidar_list = sorted(glob.glob(search_string))
lidar_pose_path = os.path.join(seq, "lidar", "poses.json")
lidar_pose = read_json(lidar_pose_path)
for idx, lidar_path in enumerate(lidar_list):
# Get pose of the lidar
translation = lidar_pose[idx]["position"]
translation = np.asarray([translation[key] for key in translation])
rotation = lidar_pose[idx]["heading"]
rotation = np.asarray([rotation[key] for key in rotation])
rotation = quaternion_to_euler(*rotation)
Rt = to_transform_matrix(translation, rotation)
# Get respective bboxes
bbox_path = lidar_path.split("/")
bbox_path[-2] = "annotations/cuboids"
bbox_path = os.path.join(*bbox_path)
# Load data
lidar = np.asarray(pd.read_pickle(lidar_path))
# Get only lidar 0 (there is also lidar 1)
lidar = lidar[lidar[:, -1] == 0]
intensity = lidar[:, 3]
lidar = transform_lidar_box_3d(lidar, Rt)
# add intensity
lidar = np.concatenate((lidar, intensity[:, None]), axis=-1)
# Load bboxes
bboxes = np.asarray(pd.read_pickle(bbox_path))
labels, bboxes = make_xzyhwly(bboxes)
corners_3d, orientation_3d = make_eight_points_boxes(bboxes)
corners_3d = np.asarray(
[transform_lidar_box_3d(box, Rt) for box in corners_3d])
orientation_3d = np.asarray(
[transform_lidar_box_3d(box, Rt) for box in orientation_3d])
labels, corners_3d, orientation_3d = filter_boxes(
labels, corners_3d, orientation_3d, lidar)
centroid, width, length, height, yaw = get_bboxes_parameters_from_points(
corners_3d)
boxes_new = np.concatenate(
(
centroid,
length[:, None],
width[:, None],
height[:, None],
yaw[:, None],
),
axis=-1,
)
lidar[:, :3] = lidar[:, :3] + shift_lidar
corners_3d, orientation_3d = make_eight_points_boxes(boxes_new)
corners_3d = corners_3d + shift_lidar
orientation_3d = orientation_3d + shift_lidar
figure = visualize_bboxes_3d(corners_3d, None, orientation_3d)
figure = visualize_lidar(lidar, figure)
mlab.show(1)
input()
mlab.close(figure)
def preprocess_data(dataset_dir):
"""
The function visualizes data from pandaset.
Arguments:
dataset_dir: directory with Pandaset data
"""
# Get list of data samples
search_string = os.path.join(dataset_dir, "*")
seq_list = sorted(glob.glob(search_string))
for seq in tqdm(seq_list, desc="Process sequences", total=len(seq_list)):
# Make output dirs for data
lidar_out_dir = os.path.join(seq, "lidar_processed")
bbox_out_dir = os.path.join(seq, "bbox_processed")
os.makedirs(lidar_out_dir, exist_ok=True)
os.makedirs(bbox_out_dir, exist_ok=True)
search_string = os.path.join(seq, "lidar", "*.pkl.gz")
lidar_list = sorted(glob.glob(search_string))
lidar_pose_path = os.path.join(seq, "lidar", "poses.json")
lidar_pose = read_json(lidar_pose_path)
for idx, lidar_path in enumerate(lidar_list):
sample_idx = os.path.splitext(os.path.basename(lidar_path))[0].split(".")[0]
# Get pose of the lidar
translation = lidar_pose[idx]["position"]
translation = np.asarray([translation[key] for key in translation])
rotation = lidar_pose[idx]["heading"]
rotation = np.asarray([rotation[key] for key in rotation])
rotation = quaternion_to_euler(*rotation)
Rt = to_transform_matrix(translation, rotation)
# Get respective bboxes
bbox_path = lidar_path.split("/")
bbox_path[-2] = "annotations/cuboids"
bbox_path = os.path.join(*bbox_path)
# Load data
lidar = np.asarray(pd.read_pickle(lidar_path))
# Get only lidar 0 (there is also lidar 1)
lidar = lidar[lidar[:, -1] == 0]
intensity = lidar[:, 3]
lidar = transform_lidar_box_3d(lidar, Rt)
# add intensity
lidar = np.concatenate((lidar, intensity[:, None]), axis=-1)
# Load bboxes
bboxes = np.asarray(pd.read_pickle(bbox_path))
labels, bboxes = make_xzyhwly(bboxes)
corners_3d, orientation_3d = make_eight_points_boxes(bboxes)
corners_3d = np.asarray(
[transform_lidar_box_3d(box, Rt) for box in corners_3d]
)
orientation_3d = np.asarray(
[transform_lidar_box_3d(box, Rt) for box in orientation_3d]
)
# filter boxes containing less then 20 lidar points inside
labels, corners_3d, orientation_3d = filter_boxes(
labels, corners_3d, orientation_3d, lidar
)
centroid, width, length, height, yaw = get_bboxes_parameters_from_points(
corners_3d
)
# Save data
lidar_filename = os.path.join(lidar_out_dir, sample_idx + ".bin")
save_lidar(lidar_filename, lidar.astype(np.float32))
box_filename = os.path.join(bbox_out_dir, sample_idx + ".txt")
save_bboxes_to_file(
box_filename, centroid, width, length, height, yaw, labels
)
def get_report_json(
checkpoints_dir, report_dir, plot_metrics, main_metric, full_report=False
):
"""
Parse checkpoints_dir and returns json report with all information
Arguments:
checkpoints_dir: ditectory containing all checkpoints from training
report_dir: directory to save reprot
plot_metrics: metrics to plot
main_metric: main metric to define best epoch
full_report: generate .json file containing all information
from all epoch (not only from the best)
Returns:
report_dict: the dictionary with information for report
"""
search_string = os.path.join(checkpoints_dir, "*")
model_list = sorted(glob.glob(search_string))
date_now = datetime.datetime.now().strftime("%d %B %Y")
report_dict = {
"model_name": None,
"date": date_now,
"parameters": None,
"best_epoch": None,
"main_metric": main_metric,
"epoch_metrics": None,
"plot_metrics": None,
}
if len(model_list) == 0:
ValueError("The checkpoint folder {} is empty".format(checkpoints_dir))
else:
epoch_metrics = {}
for model_folder in model_list:
model_name, epoch = model_folder.split("/")[-1].split("-")
# Fill header
if report_dict["model_name"] is None and report_dict["parameters"] is None:
param_filename = os.path.join(model_folder, "parameters.json")
report_dict["model_name"] = model_name
report_dict["parameters"] = read_json(param_filename)
# Check that we have only one model name inside checkpoints
if model_name != report_dict["model_name"]:
ValueError(
"model name in report {} is not \
same as current model folder {}".format(
report_dict["model_name"], model_name
)
)
# Fill epoch metrics
metrics_filename = os.path.join(model_folder, "epoch_metrics.json")
epoch_metrics[epoch] = read_json(metrics_filename)
# Find best epoch by metric
report_dict["epoch_metrics"] = epoch_metrics
# Get metrics to plot
plot_metrics = parse_report_for_epoch_metrics(report_dict, metrics=plot_metrics)
# Find best checkpoint idx, epoch
best_ckpt_idx = np.argmax(plot_metrics[main_metric]["value"])
best_epoch = "{0:04d}".format(plot_metrics[main_metric]["epoch"][best_ckpt_idx])
report_dict["best_epoch"] = best_epoch
report_dict["plot_metrics"] = plot_metrics
if not full_report:
# Override the epoch metrics with info from only best epoch
report_dict["epoch_metrics"] = {best_epoch: epoch_metrics[best_epoch]}
return report_dict
from detection_3d.tools.visualization_tools import visualize_lidar, visualize_bboxes_3d import mayavi.mlab as mlab import numpy as np from detection_3d.tools.file_io import read_json param_settings = Parameters().settings from detection_3d.reproject_to_image import intrinscs_to_matrix, get_extrinsic bboxes_file = "/home/denis/lidar_dynamic_objects_detection/detection_3d/dataset/001/bbox_processed/00.txt" lidar_file = "/home/denis/Pandaset/PandaSet/001/lidar_processed/00.bin" pose_lidar = "/home/denis/Pandaset/PandaSet/001/lidar/poses.json" pose_camera = "/home/denis/Pandaset/PandaSet/001/camera/front_camera/poses.json" intrinsics = "/home/denis/Pandaset/PandaSet/001/camera/front_camera/intrinsics.json" pose_l = read_json(pose_lidar) pose_c = read_json(pose_camera) intr_c = read_json(intrinsics) cam_T_lid = get_extrinsic(pose_l[0], pose_c[0]) lidar = load_lidar(lidar_file) lidar = lidar[:, :3] lidar = np.transpose(lidar) ones = np.ones_like(lidar[0]) lidar_hom = np.concatenate((lidar, ones[None, :]), axis=0) lidar_cam = np.matmul(cam_T_lid, lidar_hom) lidar_cam = np.transpose(lidar_cam) lidar = lidar_cam[:, :3] lidar = lidar[lidar[:, 2] >= 0] bboxes = load_bboxes(bboxes_file)
