s23dr-hoho-competition / handcrafted_solution.py
Siromanec's picture
added notebooks
dd68072
raw history blame
No virus
34.4 kB
# Description: This file contains the handcrafted solution for the task of wireframe reconstruction
import io
from collections import defaultdict
from typing import Tuple, List
import cv2
import hoho
import numpy as np
import scipy.interpolate as si
from PIL import Image as PImage
from hoho.color_mappings import gestalt_color_mapping
from hoho.read_write_colmap import read_cameras_binary, read_images_binary, read_points3D_binary
from scipy.spatial import KDTree
from scipy.spatial.distance import cdist
from sklearn.cluster import DBSCAN
from scipy.spatial import cKDTree
from enum import Enum
apex_color = gestalt_color_mapping["apex"]
eave_end_point = gestalt_color_mapping["eave_end_point"]
flashing_end_point = gestalt_color_mapping["flashing_end_point"]
apex_color, eave_end_point, flashing_end_point = [np.array(i) for i in [apex_color, eave_end_point, flashing_end_point]]
unclassified = np.array([(215, 62, 138)])
line_classes = ['eave', 'ridge', 'rake', 'valley']
class VertexType(Enum):
APEX = 0
EAVE_END_POINT = 1
class NearestNDInterpolatorWithThreshold(si.NearestNDInterpolator):
def __init__(self, points, values, max_distance):
super().__init__(points, values)
self.max_distance = max_distance
self.tree = cKDTree(points)
def __call__(self, *args):
# Convert the input to a 2D array of query points
query_points = np.array(args).T
distances, indices = self.tree.query(query_points, k=5, distance_upper_bound=self.max_distance)
# valid_mask = distances <= self.max_distance
# print(indices)
# print(distances)
# distances = np.mean(distances, axis=1)
# print(indices)
# print(self.values)# print(distances)
found_mask = indices != len(self.values)
temp_values = np.concatenate([self.values, [0]])
values = temp_values[indices]
# values[~found_mask] = 0
values = np.sum(values, axis=1)
found_mask_sum = np.sum(found_mask, axis=1)
found_mask = found_mask_sum != 0
values[found_mask] /= found_mask_sum[found_mask]
values[~found_mask] = np.nan
# print(values)
# np.ma.masked_where(indices, found_mask)
# print(self.values[indices[found_mask]])
# values[found_mask] =
# print(values)
return values.T
def empty_solution():
'''Return a minimal valid solution, i.e. 2 vertices and 1 edge.'''
return np.zeros((2, 3)), [(0, 1)]
def convert_entry_to_human_readable(entry):
out = {}
already_good = {'__key__', 'wf_vertices', 'wf_edges', 'edge_semantics', 'mesh_vertices', 'mesh_faces',
'face_semantics', 'K', 'R', 't'}
for k, v in entry.items():
if k in already_good:
out[k] = v
continue
match k:
case 'points3d':
out[k] = read_points3D_binary(fid=io.BytesIO(v))
case 'cameras':
out[k] = read_cameras_binary(fid=io.BytesIO(v))
case 'images':
out[k] = read_images_binary(fid=io.BytesIO(v))
case 'ade20k' | 'gestalt':
out[k] = [PImage.open(io.BytesIO(x)).convert('RGB') for x in v]
case 'depthcm':
out[k] = [PImage.open(io.BytesIO(x)) for x in entry['depthcm']]
return out
def remove_undesired_objects(image):
image = image.astype('uint8')
nb_components, output, stats, centroids = cv2.connectedComponentsWithStats(image, connectivity=4)
sizes = stats[:, -1]
max_label = 1
max_size = sizes[1]
for i in range(2, nb_components):
if sizes[i] > max_size:
max_label = i
max_size = sizes[i]
img2 = np.zeros(output.shape)
img2[output == max_label] = 1
return img2
def clean_image(image_gestalt) -> np.ndarray:
# clears image in from of unclassified and disconected components
image_gestalt = np.array(image_gestalt)
# unclassified_mask = cv2.inRange(image_gestalt, unclassified - 1, unclassified + 1)
# unclassified_mask = cv2.bitwise_not(unclassified_mask)
# mask = remove_undesired_objects(unclassified_mask).astype(np.uint8)
# mask = cv2.morphologyEx(mask, cv2.MORPH_CLOSE, np.ones((11, 11), np.uint8), iterations=11)
# mask = cv2.morphologyEx(mask, cv2.MORPH_DILATE, np.ones((11, 11), np.uint8), iterations=2)
# image_gestalt[:, :, 0] *= mask
# image_gestalt[:, :, 1] *= mask
# image_gestalt[:, :, 2] *= mask
return image_gestalt
def get_vertices(image_gestalt, *, color_range=3.5, dialations=2, erosions=1, kernel_size=11):
### detects the apex and eave end and flashing end points
apex_mask = cv2.inRange(image_gestalt, apex_color - color_range, apex_color + color_range)
eave_end_point_mask = cv2.inRange(image_gestalt, eave_end_point - color_range, eave_end_point + color_range)
flashing_end_point_mask = cv2.inRange(image_gestalt, flashing_end_point - color_range,
flashing_end_point + color_range)
eave_end_point_mask = cv2.bitwise_or(eave_end_point_mask, flashing_end_point_mask)
kernel = np.ones((kernel_size, kernel_size), np.uint8)
apex_mask = cv2.morphologyEx(apex_mask, cv2.MORPH_DILATE, kernel, iterations=dialations)
apex_mask = cv2.morphologyEx(apex_mask, cv2.MORPH_ERODE, kernel, iterations=erosions)
eave_end_point_mask = cv2.morphologyEx(eave_end_point_mask, cv2.MORPH_DILATE, kernel, iterations=dialations)
eave_end_point_mask = cv2.morphologyEx(eave_end_point_mask, cv2.MORPH_ERODE, kernel, iterations=erosions)
*_, apex_stats, apex_centroids = cv2.connectedComponentsWithStats(apex_mask, connectivity=4, stats=cv2.CV_32S)
*_, other_stats, other_centroids = cv2.connectedComponentsWithStats(eave_end_point_mask, connectivity=4, stats=cv2.CV_32S)
return (apex_centroids[1:],
other_centroids[1:],
apex_mask,
eave_end_point_mask,
np.maximum(apex_stats[1:, cv2.CC_STAT_WIDTH], apex_stats[1:, cv2.CC_STAT_HEIGHT])/2,
np.maximum(other_stats[1:, cv2.CC_STAT_WIDTH], other_stats[1:, cv2.CC_STAT_HEIGHT])/2)
def infer_vertices(image_gestalt, *, color_range=4.):
ridge_color = np.array(gestalt_color_mapping["ridge"])
rake_color = np.array(gestalt_color_mapping["rake"])
ridge_mask = cv2.inRange(image_gestalt,
ridge_color - color_range,
ridge_color + color_range)
ridge_mask = cv2.morphologyEx(ridge_mask,
cv2.MORPH_DILATE, np.ones((3, 3)), iterations=4)
rake_mask = cv2.inRange(image_gestalt,
rake_color - color_range,
rake_color + color_range)
rake_mask = cv2.morphologyEx(rake_mask,
cv2.MORPH_DILATE, np.ones((3, 3)), iterations=4)
intersection_mask = cv2.bitwise_and(ridge_mask, rake_mask)
intersection_mask = cv2.morphologyEx(intersection_mask, cv2.MORPH_DILATE, np.ones((11, 11)), iterations=3)
*_, inferred_centroids = cv2.connectedComponentsWithStats(intersection_mask, connectivity=4, stats=cv2.CV_32S)
return inferred_centroids[1:], intersection_mask
def get_missed_vertices(vertices, inferred_centroids, *, min_missing_distance=200.0, **kwargs):
vertices = KDTree(vertices)
closest = vertices.query(inferred_centroids, k=1, distance_upper_bound=min_missing_distance)
missed_points = inferred_centroids[closest[1] == len(vertices.data)]
return missed_points
def get_lines_and_directions(gest_seg_np, edge_class, *, color_range=4., rho, theta, threshold, min_line_length,
max_line_gap, extend=30, kernel_size=3, dilation_iterations=1, **kwargs):
edge_color = np.array(gestalt_color_mapping[edge_class])
mask = cv2.inRange(gest_seg_np,
edge_color - color_range,
edge_color + color_range)
mask = cv2.morphologyEx(mask,
cv2.MORPH_DILATE, np.ones((kernel_size, kernel_size)), iterations=dilation_iterations)
if not np.any(mask):
return [], []
# Run Hough on edge detected image
# Output "lines" is an array containing endpoints of detected line segments
# cv2.GaussianBlur(mask, (11, 11), 0, mask)
lines = cv2.HoughLinesP(mask, rho, theta, threshold, np.array([]),
min_line_length, max_line_gap)
if lines is None:
return [], []
line_directions = []
edges = []
for line_idx, line in enumerate(lines):
for x1, y1, x2, y2 in line:
if x1 < x2:
x1, y1, x2, y2 = x2, y2, x1, y1
direction = (np.array([x2 - x1, y2 - y1]))
direction = direction / np.linalg.norm(direction)
for extend_value in range(0, int(extend), 5):
new_direction = extend_value * direction
x1, y1 = -new_direction + (x1, y1)
x2, y2 = + new_direction + (x2, y2)
line_directions.append(direction)
edges.append((x1, y1, x2, y2))
return edges, line_directions
def infer_missing_vertices(ridge_edges, rake_edges):
ridge_edges = np.array(ridge_edges)
rake_edges = np.array(rake_edges)
ridge_ends = np.concatenate([ridge_edges[:, 2:], ridge_edges[:, :2]])
rake_ends = np.concatenate([rake_edges[:, 2:], rake_edges[:, :2]])
ridge_ends = KDTree(ridge_ends)
rake_ends = KDTree(rake_ends)
missing_candidates = rake_ends.query_ball_tree(ridge_ends, 10)
missing_candidates = np.concatenate([*missing_candidates])
missing_candidates = np.unique(missing_candidates).astype(np.int32)
return ridge_ends.data[missing_candidates]
def get_vertices_and_edges_from_segmentation(gest_seg_np, *,
point_radius=30,
max_angle=5.,
point_radius_scale=1,
**kwargs):
'''Get the vertices and edges from the gestalt segmentation mask of the house'''
# Apex
connections = []
deviation_threshold = np.cos(np.deg2rad(max_angle))
(apex_centroids, eave_end_point_centroids,
apex_mask, eave_end_point_mask,
apex_radii, eave_radii) = get_vertices(gest_seg_np)
vertices = np.concatenate([apex_centroids, eave_end_point_centroids])
# inferred_vertices, inferred_mask = infer_vertices(gest_seg_np)
# missed_vertices = get_missed_vertices(vertices, inferred_vertices, **kwargs)
# vertices = np.concatenate([vertices, missed_vertices])
edges = []
line_directions = []
rho = 1 # distance resolution in pixels of the Hough grid
theta = np.pi / 180 # angular resolution in radians of the Hough grid
threshold = 40 # minimum number of votes (intersections in Hough grid cell)
min_line_length = 60 # minimum number of pixels making up a line
max_line_gap = 40 # maximum gap in pixels between connectable line segments
ridge_edges, ridge_directions = get_lines_and_directions(gest_seg_np, "ridge",
rho=rho,
theta=theta,
threshold=threshold,
min_line_length=min_line_length,
max_line_gap=max_line_gap,
kernel_size=3,
dilation_iterations=3,
**kwargs)
rake_edges, rake_directions = get_lines_and_directions(gest_seg_np, "rake",
rho=rho,
theta=theta,
threshold=threshold,
min_line_length=min_line_length,
max_line_gap=max_line_gap,
**kwargs)
if len(ridge_edges) > 0:
edges.append(ridge_edges)
line_directions.append(ridge_directions)
if len(rake_edges) > 0:
edges.append(rake_edges)
line_directions.append(rake_directions)
missed_vertices = []
if len(ridge_edges) > 0 and len(rake_edges) > 0:
inferred_vertices = infer_missing_vertices(ridge_edges, rake_edges)
missed_vertices = get_missed_vertices(vertices, inferred_vertices, **kwargs)
vertices = np.concatenate([vertices, missed_vertices])
if len(vertices) < 2:
return [], []
vertex_size = np.full(len(vertices), point_radius/2)
if len(apex_radii) > 0 and len(eave_radii) > 0:
apex_radii *= point_radius_scale
eave_radii *= point_radius_scale
apex_radii = np.maximum(apex_radii, 10)
eave_radii = np.maximum(eave_radii, 10)
point_radius = np.max([np.max(apex_radii), np.max(eave_radii)])
vertex_size[:len(apex_radii)] = apex_radii
vertex_size[len(apex_radii):len(apex_radii) + len(eave_radii)] = eave_radii
# for i, coords in enumerate(vertices):
# coords = np.round(coords).astype(np.uint32)
# radius = point_radius # np.clip(int(max_depth//2 + depth_np[coords[1], coords[0]]), 10, 30)#int(np.clip(max_depth - depth_np[coords[1], coords[0]], 10, 20))
# vertex_size[i] = scale * radius
vertices = KDTree(vertices)
for edge_class in ['eave',
'step_flashing',
'flashing',
# 'post',
'valley',
'hip',
'transition_line',
'fascia',
'soffit',]:
class_edges, class_directions = get_lines_and_directions(gest_seg_np, edge_class,
rho=rho,
theta=theta,
threshold=threshold,
min_line_length=min_line_length,
max_line_gap=max_line_gap,
**kwargs)
if len(class_edges) > 0:
edges.append(class_edges)
line_directions.append(class_directions)
edges = np.concatenate(edges).astype(np.float64)
if len(edges) < 1:
return [], []
line_directions = np.concatenate(line_directions).astype(np.float64)
# calculate the distances between the vertices and the edge ends
begin_edges = KDTree(edges[:, :2])
end_edges = KDTree(edges[:, 2:])
begin_indices = begin_edges.query_ball_tree(vertices, point_radius)
end_indices = end_edges.query_ball_tree(vertices, point_radius)
line_indices = np.where(np.array([len(i) and len(j) for i, j in zip(begin_indices, end_indices)]))[0]
# create all possible connections between begin and end candidates that correspond to a line
begin_vertex_list = []
end_vertex_list = []
line_idx_list = []
for line_idx in line_indices:
begin_vertices, end_vertices = begin_indices[line_idx], end_indices[line_idx]
begin_vertices, end_vertices = np.array(begin_vertices), np.array(end_vertices)
begin_value = begin_edges.data[line_idx]
end_value = end_edges.data[line_idx]
begin_in_range_indices = np.where(
np.linalg.norm(vertices.data[begin_vertices] - begin_value, axis=1)
<
vertex_size[begin_vertices])[0]
end_in_range_indices = np.where(
np.linalg.norm(vertices.data[end_vertices] - end_value, axis=1)
<
vertex_size[end_vertices])[0]
begin_vertices = begin_vertices[begin_in_range_indices]
end_vertices = end_vertices[end_in_range_indices]
if len(begin_vertices) < 1 or len(end_vertices) < 1:
continue
begin_vertices, end_vertices = np.meshgrid(begin_vertices, end_vertices)
begin_vertex_list.extend(begin_vertices.flatten())
end_vertex_list.extend(end_vertices.flatten())
line_idx_list.extend([line_idx] * len(begin_vertices.flatten()))
line_idx_list = np.array(line_idx_list)
all_connections = np.array([begin_vertex_list, end_vertex_list])
# decrease the number of possible connections to reduce number of calculations
possible_connections = np.unique(all_connections, axis=1)
possible_connections = np.sort(possible_connections, axis=0)
possible_connections = np.unique(possible_connections, axis=1)
possible_connections = possible_connections[:, possible_connections[0, :] != possible_connections[1, :]]
if possible_connections.shape[1] < 1:
return [], []
# precalculate the possible direction vectors
possible_direction_vectors = vertices.data[possible_connections[0]] - vertices.data[possible_connections[1]]
possible_direction_vectors = possible_direction_vectors / np.linalg.norm(possible_direction_vectors, axis=1)[:,
np.newaxis]
owned_lines_per_possible_connections = [list() for i in range(possible_connections.shape[1])]
# assign lines to possible connections
for line_idx, i, j in zip(line_idx_list, begin_vertex_list, end_vertex_list):
if i == j:
continue
i, j = min(i, j), max(i, j)
for connection_idx, connection in enumerate(possible_connections.T):
if np.all((i, j) == connection):
owned_lines_per_possible_connections[connection_idx].append(line_idx)
break
# check if the lines are in the same direction as the possible connection
for fitted_line_idx, owned_lines_per_possible_connection in enumerate(owned_lines_per_possible_connections):
line_deviations = np.abs(
np.dot(line_directions[owned_lines_per_possible_connection], possible_direction_vectors[fitted_line_idx]))
if np.any(line_deviations > deviation_threshold):
connections.append(possible_connections[:, fitted_line_idx])
vertices = [{"xy": v, "type": VertexType.APEX} for v in apex_centroids]
vertices += [{"xy": v, "type": VertexType.APEX} for v in missed_vertices]
vertices += [{"xy": v, "type": VertexType.EAVE_END_POINT} for v in eave_end_point_centroids]
return vertices, connections
def get_uv_depth(vertices, depth):
'''Get the depth of the vertices from the depth image'''
depth[depth > 3000] = np.nan
uv = np.array([v['xy'] for v in vertices])
uv_int = uv.astype(np.int32)
H, W = depth.shape[:2]
uv_int[:, 0] = np.clip(uv_int[:, 0], 0, W - 1)
uv_int[:, 1] = np.clip(uv_int[:, 1], 0, H - 1)
vertex_depth = depth[(uv_int[:, 1], uv_int[:, 0])]
return uv, vertex_depth
def merge_vertices_3d(vert_edge_per_image, merge_th=0.1, **kwargs):
'''Merge vertices that are close to each other in 3D space and are of same types'''
all_3d_vertices = []
connections_3d = []
cur_start = 0
types = []
for cimg_idx, (vertices, connections, vertices_3d) in vert_edge_per_image.items():
# remove nan values and remap the connections
connections = [[a, b]
for (a, b) in connections
if
not np.any(np.isnan(vertices_3d[a]))
and
not np.any(np.isnan(vertices_3d[b]))
]
left_vertex_indices = np.where(np.all(~np.isnan(vertices_3d), axis=1))[0]
new_indices = np.arange(len(left_vertex_indices))
new_vertex_mapping = dict(zip(left_vertex_indices, new_indices))
vertices = [v for i, v in enumerate(vertices) if i in new_vertex_mapping]
types += [int(v['type'] == VertexType.APEX) for v in vertices]
vertices_3d = vertices_3d[left_vertex_indices]
connections = [[new_vertex_mapping[a] + cur_start, new_vertex_mapping[b] + cur_start] for a, b in connections]
all_3d_vertices.append(vertices_3d)
connections_3d += connections
cur_start += len(vertices_3d)
all_3d_vertices = np.concatenate(all_3d_vertices, axis=0)
distmat = cdist(all_3d_vertices, all_3d_vertices)
types = np.array(types).reshape(-1, 1)
same_types = cdist(types, types)
mask_to_merge = (distmat <= merge_th) & (same_types == 0)
new_vertices = []
new_connections = []
to_merge = sorted(list(set([tuple(a.nonzero()[0].tolist()) for a in mask_to_merge])))
to_merge_final = defaultdict(list)
for i in range(len(all_3d_vertices)):
for j in to_merge:
if i in j:
to_merge_final[i] += j
for k, v in to_merge_final.items():
to_merge_final[k] = list(set(v))
already_there = set()
merged = []
for k, v in to_merge_final.items():
if k in already_there:
continue
merged.append(v)
for vv in v:
already_there.add(vv)
old_idx_to_new = {}
for count, idxs in enumerate(merged):
new_vertices.append(all_3d_vertices[idxs].mean(axis=0))
for idx in idxs:
old_idx_to_new[idx] = count
new_vertices = np.array(new_vertices)
for conn in connections_3d:
new_con = sorted((old_idx_to_new[conn[0]], old_idx_to_new[conn[1]]))
if new_con[0] == new_con[1]:
continue
if new_con not in new_connections:
new_connections.append(new_con)
return new_vertices, new_connections
def prune_not_connected(all_3d_vertices, connections_3d):
'''Prune vertices that are not connected to any other vertex'''
connected = defaultdict(list)
for c in connections_3d:
connected[c[0]].append(c)
connected[c[1]].append(c)
new_indexes = {}
new_verts = []
connected_out = []
for k, v in connected.items():
vert = all_3d_vertices[k]
if tuple(vert) not in new_verts:
new_verts.append(tuple(vert))
new_indexes[k] = len(new_verts) - 1
for k, v in connected.items():
for vv in v:
connected_out.append((new_indexes[vv[0]], new_indexes[vv[1]]))
connected_out = list(set(connected_out))
return np.array(new_verts), connected_out
def clean_points3d(entry, clustering_eps):
image_dict = {}
for k, v in entry["images"].items():
image_dict[v.name] = v
points = [v.xyz for k, v in entry["points3d"].items()]
points = np.array(points)
point_keys = [k for k, v in entry["points3d"].items()]
point_keys = np.array(point_keys)
clustered = DBSCAN(eps=clustering_eps, min_samples=5).fit(points).labels_
clustered_indices = np.argsort(clustered)
points = points[clustered_indices]
point_keys = point_keys[clustered_indices]
clustered = clustered[clustered_indices]
_, cluster_indices = np.unique(clustered, return_index=True)
clustered_points = np.split(points, cluster_indices[1:])
clustered_keys = np.split(point_keys, cluster_indices[1:])
biggest_cluster_index = np.argmax([len(i) for i in clustered_points])
biggest_cluster = clustered_points[biggest_cluster_index]
biggest_cluster_keys = clustered_keys[biggest_cluster_index]
biggest_cluster_keys = set(biggest_cluster_keys)
points3d_kdtree = KDTree(biggest_cluster)
return points3d_kdtree, biggest_cluster_keys, image_dict
def get_depth_from_pointcloud(image, pointcloud, biggest_cluster_keys, R, t):
belonging_points3d = []
belonging_points2d = []
point_indices = np.where(image.point3D_ids != -1)[0]
for idx, point_id in zip(point_indices, image.point3D_ids[point_indices]):
if point_id in biggest_cluster_keys:
belonging_points3d.append(pointcloud[point_id].xyz)
belonging_points2d.append(image.xys[idx])
if len(belonging_points3d) < 1:
print(f'No 3D points in image {image.name}')
raise KeyError
belonging_points3d = np.array(belonging_points3d)
belonging_points2d = np.array(belonging_points2d)
# projected2d, _ = cv2.projectPoints(belonging_points3d, R, t, K, dist_coeff)
important = np.where(np.all(belonging_points2d >= 0, axis=1))
# Normalize the uv to the camera intrinsics
world_to_cam = np.eye(4)
world_to_cam[:3, :3] = R
world_to_cam[:3, 3] = t
homo_belonging_points = cv2.convertPointsToHomogeneous(belonging_points3d)
depth = cv2.convertPointsFromHomogeneous(cv2.transform(homo_belonging_points, world_to_cam))
depth = depth[:, 0, 2]
# projected2d = projected2d[:, 0, :]
depth = depth[important[0]]
# projected2d = projected2d[important[0]]
projected2d = belonging_points2d[important[0]]
return projected2d, depth
def get_depthmap_from_pointcloud(pointcloud, biggest_cluster_keys, K, R, t, depthmap):
belonging_points3d = []
for point_id in biggest_cluster_keys:
belonging_points3d.append(pointcloud[point_id].xyz)
belonging_points3d = np.array(belonging_points3d)
projected2d, _ = cv2.projectPoints(belonging_points3d, R, t, K, 0)
projected2d = projected2d.reshape(-1, 2)
important = np.where(np.all(projected2d >= 0, axis=1))[0]
# Normalize the uv to the camera intrinsics
world_to_cam = np.eye(4)
world_to_cam[:3, :3] = R
world_to_cam[:3, 3] = t
homo_belonging_points = cv2.convertPointsToHomogeneous(belonging_points3d)
depth = cv2.convertPointsFromHomogeneous(cv2.transform(homo_belonging_points, world_to_cam))
depth = depth[:, 0, 2]
# projected2d = projected2d[:, 0, :]
depth = depth[important]
projected2d = projected2d[important]
projected2d = projected2d.astype(np.int32)
projected2d[:, 1] = np.clip(projected2d[:, 1], 0, depthmap.shape[1]-1)
projected2d[:, 0] = np.clip(projected2d[:, 0], 0, depthmap.shape[0]-1)
pointcloud_depthmap = np.full_like(depthmap, 4000)
sorted_indices = np.argsort(depth)[::-1]
projected2d, depth = projected2d[sorted_indices], depth[sorted_indices]
idx = np.searchsorted(depth, 1000)
projected2d, depth = projected2d[:idx], depth[:idx]
for point, depth_value in zip(projected2d, depth):
cv2.circle(pointcloud_depthmap, point, 20, depth_value, -1)
return pointcloud_depthmap
def get_mesh_depthmap(mesh, K, R, t, depthmap, fill_value=4000):
world_to_cam = np.eye(4)
world_to_cam[:3, :3] = R
world_to_cam[:3, 3] = t
camera_transform = np.linalg.inv(world_to_cam)
import trimesh
camera = trimesh.scene.Camera(focal=(K[0, 0], K[1, 1]), resolution=depthmap.shape[::-1])
scene = trimesh.scene.scene.Scene(camera=camera, camera_transform=camera_transform)
ray_origins, ray_directions, corresponding_pixels = scene.camera_rays()
locations, index_ray, index_tri = mesh.ray.intersects_location(ray_origins, -ray_directions, multiple_hits=False)
mesh_depths = np.linalg.norm(ray_origins[index_ray] - locations, axis=1)
mesh_depthmap = np.full(camera.resolution, fill_value)
mesh_depthmap[corresponding_pixels[index_ray, 0], corresponding_pixels[index_ray, 1]] = mesh_depths
mesh_depthmap = cv2.flip(cv2.rotate(mesh_depthmap, cv2.ROTATE_90_COUNTERCLOCKWISE), 1)
return mesh_depthmap
def predict(entry, visualize=False,
scale_estimation_coefficient=2.5,
clustering_eps=100,
dist_coeff=0,
pointcloud_depth_coeff = 1,
interpolation_radius=200,
**kwargs) -> Tuple[np.ndarray, List[int]]:
if 'gestalt' not in entry or 'depthcm' not in entry or 'K' not in entry or 'R' not in entry or 't' not in entry:
print('Missing required fields in the entry')
return (entry['__key__'], *empty_solution())
entry = hoho.decode(entry)
vert_edge_per_image = {}
points3d_kdtree, biggest_cluster_keys, image_dict = clean_points3d(entry, clustering_eps)
# import trimesh
# mesh = trimesh.Trimesh(vertices=entry["mesh_vertices"], faces=entry["mesh_faces"][:,1:], use_embree=True)
for i, (gest, depthcm, K, R, t, imagekey) in enumerate(zip(entry['gestalt'],
entry['depthcm'],
entry['K'],
entry['R'],
entry['t'],
entry['__imagekey__']
)):
gest_seg = gest.resize(depthcm.size)
gest_seg_np = np.array(gest_seg).astype(np.uint8)
vertices, connections = get_vertices_and_edges_from_segmentation(gest_seg_np, **kwargs)
if (len(vertices) < 2) or (len(connections) < 1):
print(f'Not enough vertices or connections in image {i}')
vert_edge_per_image[i] = np.empty((0, 2)), [], np.empty((0, 3))
continue
depth_np = np.array(depthcm) / scale_estimation_coefficient
# depth_np = get_mesh_depthmap(mesh, K, R, t, depth_np, 4000).astype(np.float32)
# kernel = np.array([[-1,-1,-1], [-1,9,-1], [-1,-1,-1]])
# depth_np = cv2.filter2D(depth_np, -1, kernel)
uv, depth_vert_from_depth_map = get_uv_depth(vertices, depth_np)
try:
# raise KeyError
image = image_dict[imagekey]
projected2d, depth = get_depth_from_pointcloud(image, entry["points3d"], biggest_cluster_keys, R, t)
if len(depth) < 1:
print(f'No 3D points in image {i}')
# vert_edge_per_image[i] = np.empty((0, 2)), [], np.empty((0, 3))
raise KeyError
depth *= pointcloud_depth_coeff
# interpolator = si.NearestNDInterpolator(projected2d, depth, rescale=True)
interpolator = NearestNDInterpolatorWithThreshold(projected2d, depth, interpolation_radius)
uv = np.array([v['xy'] for v in vertices])
xi, yi = uv[:, 0], uv[:, 1]
depth_vert_from_pointcloud = interpolator(xi, yi)
depthmap_used = False
except KeyError:
#Revert to the depthmap
# if len(biggest_cluster_keys) > 0:
# depth_map = get_depthmap_from_pointcloud(entry["points3d"], biggest_cluster_keys, K, R, t, depth_np)
# uv, depth_vert_from_depth_map = get_uv_depth(vertices, depth_map)
depthmap_used = True
# Normalize the uv to the camera intrinsics
xy_local = np.ones((len(uv), 3))
xy_local[:, 0] = (uv[:, 0] - K[0, 2]) / K[0, 0]
xy_local[:, 1] = (uv[:, 1] - K[1, 2]) / K[1, 1]
# Get the 3D vertices
depth_vert_nan_idxs = None
if depthmap_used:
depth_vert = depth_vert_from_depth_map
else:
depth_vert_nan_idxs = np.where(np.isnan(depth_vert_from_pointcloud))[0]
depth_vert_from_pointcloud[depth_vert_nan_idxs] = depth_vert_from_depth_map[depth_vert_nan_idxs]
depth_vert = depth_vert_from_pointcloud
norm_factor = np.linalg.norm(xy_local, axis=1)[..., None]
if depth_vert_nan_idxs is not None and len(depth_vert_nan_idxs) > 0:
norm_factor_min = np.min(norm_factor[depth_vert_nan_idxs])
if len(depth_vert_nan_idxs) != len(norm_factor):
norm_factor_max = np.max(norm_factor[~np.isin(np.arange(len(norm_factor)), depth_vert_nan_idxs)])
else:
norm_factor_max = np.max(norm_factor)
else:
norm_factor_min = np.min(norm_factor)
norm_factor_max = np.max(norm_factor)
vertices_3d_local = depth_vert[..., None] * xy_local
if depthmap_used:
vertices_3d_local /= norm_factor_max
else:
vertices_3d_local[depth_vert_nan_idxs] /= norm_factor_max
vertices_3d_local[~np.isin(np.arange(len(vertices_3d_local)), depth_vert_nan_idxs)] /= norm_factor_max
world_to_cam = np.eye(4)
world_to_cam[:3, :3] = R
world_to_cam[:3, 3] = t
cam_to_world = np.linalg.inv(world_to_cam)
vertices_3d = cv2.transform(cv2.convertPointsToHomogeneous(vertices_3d_local), cam_to_world)
vertices_3d = cv2.convertPointsFromHomogeneous(vertices_3d).reshape(-1, 3)
# if not depthmap_used:
# not_nan_items = np.all(~np.isnan(vertices_3d), axis=1)
# _, closest_fitted = points3d_kdtree.query(vertices_3d[not_nan_items])
#
# vertices_3d[not_nan_items] = points3d_kdtree.data[closest_fitted]
vert_edge_per_image[i] = vertices, connections, vertices_3d
all_3d_vertices, connections_3d = merge_vertices_3d(vert_edge_per_image, **kwargs)
all_3d_vertices_clean, connections_3d_clean = all_3d_vertices, connections_3d
# all_3d_vertices_clean, connections_3d_clean = prune_not_connected(all_3d_vertices, connections_3d)
# highest_edges = np.argpartition(all_3d_vertices_clean[:, 1], 4)[:4].tolist()
#
# connections_3d_clean.append(highest_edges[:2])
# connections_3d_clean.append(highest_edges[2:])
# all_3d_vertices_clean, connections_3d_clean = prune_not_connected(all_3d_vertices, connections_3d)
if (len(all_3d_vertices_clean) < 2) or len(connections_3d_clean) < 1:
print(f'Not enough vertices or connections in the 3D vertices')
return (entry['__key__'], *empty_solution())
if visualize:
from hoho.viz3d import plot_estimate_and_gt
plot_estimate_and_gt(all_3d_vertices_clean,
connections_3d_clean,
entry['wf_vertices'],
entry['wf_edges'])
return entry['__key__'], all_3d_vertices_clean, connections_3d_clean