import copy
from mytrain import *
import torch
import torchvision.models as models
from matplotlib import pyplot as plt
import numpy as np
import open3d as o3d
from sklearn.cluster import KMeans
from sklearn.preprocessing import StandardScaler
from mpl_toolkits.mplot3d import Axes3D
import matplotlib.pyplot as plt
import open3d as o3d
# from utils.GNSS_SPP.test import coordinateTrans
import numpy as np
import open3d as o3d
import copy
from matplotlib import pyplot as plt
# 在点云上添加分类标签
def draw_labels_on_model(pcl, labels):
cmap = plt.get_cmap("tab20")
pcl_temp = copy.deepcopy(pcl)
max_label = labels.max()
colors = cmap(labels / (max_label if max_label > 0 else 1))
pcl_temp.colors = o3d.utility.Vector3dVector(colors[:, :3])
o3d.visualization.draw_geometries([pcl_temp], window_name="可视化分类结果",
width=800, height=800, left=50, top=50,
mesh_show_back_face=False)
# 计算欧氏距离
def euclidean_distance(one_sample, X):
# 将one_sample转换为一纬向量
one_sample = one_sample.reshape(1, -1)
# 把X转换成一维向量
X = X.reshape(X.shape[0], -1)
# 这是用来确保one_sample的尺寸与X相同
distances = np.power(np.tile(one_sample, (X.shape[0], 1)) - X, 2).sum(axis=1)
return distances
class Kmeans(object):
# 构造函数
def __init__(self, k=2, max_iterations=1500, tolerance=0.00001):
self.k = k
self.max_iterations = max_iterations
self.tolerance = tolerance
# 随机选取k个聚类中心点
def init_random_centroids(self, X):
# save the shape of X
n_samples, n_features = np.shape(X)
# make a zero matrix to store values
centroids = np.zeros((self.k, n_features))
# 因为有k个中心点，所以执行k次循环
for i in range(self.k):
# 随机选取范围内的值
centroid = X[np.random.choice(range(n_samples))]
centroids[i] = centroid
return centroids
# 查找距离样本点最近的中心
def closest_centroid(self, sample, centroids):
distances = euclidean_distance(sample, centroids)
# np.argmin 返回距离最小值的下标
closest_i = np.argmin(distances)
return closest_i
# 确定聚类
def create_clusters(self, centroids, X):
# 这是为了构造用于存储集群的嵌套列表
clusters = [[] for _ in range(self.k)]
for sample_i, sample in enumerate(X):
centroid_i = self.closest_centroid(sample, centroids)
clusters[centroid_i].append(sample_i)
return clusters
# 基于均值算法更新质心
def update_centroids(self, clusters, X):
n_features = np.shape(X)[1]
centroids = np.zeros((self.k, n_features))
for i, cluster in enumerate(clusters):
centroid = np.mean(X[cluster], axis=0)
centroids[i] = centroid
return centroids
# 获取标签
def get_cluster_labels(self, clusters, X):
y_pred = np.zeros(np.shape(X)[0])
for cluster_i, cluster in enumerate(clusters):
for sample_i in cluster:
y_pred[sample_i] = cluster_i
return y_pred
# 预测标签
def predict(self, X):
# 随机选取中心点
centroids = self.init_random_centroids(X)
for _ in range(self.max_iterations):
# 对所有点进行聚类
clusters = self.create_clusters(centroids, X)
former_centroids = centroids
# 计算新的聚类中心
centroids = self.update_centroids(clusters, X)
# 判断是否满足收敛
diff = centroids - former_centroids
if diff.any() < self.tolerance:
break
return self.get_cluster_labels(clusters, X)
# if __name__ == "__main__":
# # 加载点云
# pcd = o3d.io.read_point_cloud('cluster1.pcd')
# points = np.asarray(pcd.points)
# o3d.visualization.draw_geometries([pcd], window_name="可视化原始点云",
# width=800, height=800, left=50, top=50,
# mesh_show_back_face=False)
# # 执行K-means聚类
# clf = Kmeans(k=3)
# labels = clf.predict(points)
# # 可视化聚类结果
# draw_labels_on_model(pcd, labels)
def render_signal_t(tags, rays_o, near, far, N_samples, **kwargs):
"""_summary_
Parameters
----------
tags: [batchsize, 3]. The position of tags
rays_o : [batchsize, 3]. The origin of rays
rays_d : [batchsize, 3*360*60]. The direction of rays
near : float. The near bound of the rays
far : float. The far bound of the rays
n_samples: int. num of samples per ray
"""
batchsize = tags.shape[0]
# rays_d = tr.repeat_interleave(rays_d, batchsize)
rays_d = tags - rays_o
rays_dc = rays_d.cpu()
rays_norm = tr.tensor(np.linalg.norm(rays_dc, ord=None, axis=1))
rays_norm = tr.reshape(rays_norm, (batchsize, -1, 1))
rays_d = (rays_dc / rays_norm).to(device)
rays_d = tr.reshape(rays_d, (batchsize, -1, 3))
# rays_d = tr.reshape(rays_d, (batchsize, -1, 3)) # [batchsize, 180*30, 2]
chunks = 1
chunks_num = 1
rays_o_chunks = rays_o.expand(chunks, -1, -1).permute(1, 0, 2) # [bs, cks, 3]
tags_chunk = tags.expand(chunks, -1, -1).permute(1, 0, 2) # [bs, cks, 3]
recv_signal = tr.zeros(batchsize).to(device)
for i in range(chunks_num):
rays_d_chunks = rays_d[:, i * chunks:(i + 1) * chunks, :] # [bs, 180*30, 2]
pts, t_vals = get_points(rays_o_chunks, rays_d_chunks, 0, 1, 50) # [bs, cks, pts, 3]
views_chunks = rays_d_chunks[..., None, :].expand(pts.shape) # [bs, cks, pts, 3]
tags_chunks = tags_chunk[..., None, :].expand(pts.shape) # [bs, cks, pts, 3]
raw = run_network_t(pts, views_chunks, tags_chunks, **kwargs) # [bs, cks, pts, 3]
alphac = raw[..., 0]
s = raw[..., 1]
return pts, alphac, s # [batchsize,]
def run_network_t(pts, views, tags, network_fn, embed_pts_fn, embed_view_fn, embed_tag_fn):
pts_shapes = list(pts.shape) # [batchsize,chunk,n_samples, 3]
pts = tr.reshape(pts, [-1, pts.shape[-1]]) # [batchsize*chunk*n_samples, 3]
embedded = embed_pts_fn(pts) # [batchsize*chunk*n_samples, 3*20]
views = tr.reshape(views, [-1, views.shape[-1]]).to(device)
embedded = tr.cat([embedded, embed_view_fn(views)], -1)
tags = tr.reshape(tags, [-1, tags.shape[-1]])
embedded = tr.cat([embedded, embed_tag_fn(tags)], -1)
outputs = network_fn(embedded)
outputs = tr.reshape(outputs, pts_shapes[:-1] + [outputs.shape[-1]]) # [batchsize,chunk,n_samples,4]
return outputs
def getAllAngle_t():
r_d = []
r_dp = []
beta_r = [0, 90]
beta_len = int(beta_r[1] - beta_r[0])
frac = 10
radius = 1
alphas = np.linspace(0, 360 - 1, 360 // frac) / 180 * np.pi
betas = np.linspace(beta_r[0], beta_r[1] - 1, beta_len // frac) / 180 * np.pi
print(betas)
angleLen = len(alphas) * len(betas)
for i in range(len(alphas)):
for j in range(len(betas)):
r_d.append([alphas[i], betas[j]])
for k in range(angleLen):
x = radius * np.cos(r_d[k][0]) * np.cos(r_d[k][1]) # (90*360)
y = radius * np.sin(r_d[k][0]) * np.cos(r_d[k][1])
z = radius * np.sin(r_d[k][1])
r_dp.append([x, y, z])
return tr.tensor(r_dp)
def countOcc(xc, yc, zc):
xn = []
yn = []
zn = []
n = -40
for index in xc:
if index > 6:
xn.append(n)
n = n + 1
n = -40
for index in yc:
if index > 6:
yn.append(n)
n = n + 1
n = -40
for index in zc:
if index > 7:
zn.append(n)
n = n + 1
return xn, yn, zn
def getSPDensity():
r = 6378137
# xNum = 40
# yNum = 50
# zNum = 40
# x_nSamples = np.linspace(37.5295853 ,37.5306248, xNum)
# y_nSamples = np.linspace(122.0717183, 122.0738229, yNum)
# z_nSamples = np.linspace(15.213745 ,100, zNum)
xNum = 35
yNum = 35
zNum = 35
x_nSamples = np.linspace(0, 1, xNum)
y_nSamples = np.linspace(0, 1, yNum)
z_nSamples = np.linspace(0, 0.25 , zNum)
# y_Nsamples = np.linspace()
SP = []
for x in x_nSamples:
for y in y_nSamples:
for z in z_nSamples:
# x1, y1, z1 = coordinateTrans(x, y, z)
# x1 = x1 / (2 * r)
# y1 = y1 / (2 * r)
# z1 = z1 / (2 * r)
# SP.append([x1, y1, z1])
SP.append([x, y, z])
print(x, y, z)
# SP = tr.tensor(SP)
# tr.reshape(SP, (-1, 3))
# views_chunks = tr.tensor([0, 0, 0]).expand(SP.shape).to(device)
# tags_chunks = tr.tensor([0, 0, 0]).expand(SP.shape).to(device)
#
# raw = run_network(SP, views_chunks, tags_chunks, **kwargs) # [bs, cks, pts, 3]
# alphac = raw[..., 0]
# print(alphac)
# s = raw[..., 1]
# # print(alphac.shape)
return SP
if __name__ == '__main__':
fPone = open("data/pointCloud/pp.txt", "w+", encoding='utf-8')
fPosition = open("data/pointCloud/position.txt", "w+", encoding='utf-8')
fS = open("data/pointCloud/sPosition.txt", "w+", encoding='utf-8')
dt = pd.read_csv("data/my-exp4/dataTrain.csv")
data = dt.values
parser = config_parser()
args = parser.parse_args()
render_kwargs_train, render_kwargs_train2, render_kwargs_test, start, grad_vars, optimizer = create_nerf(args)
batchdata = tr.FloatTensor(data[200:2200].astype(float))
pPosition = batchdata[..., 0:3].cpu()
# pPosition = batchdata[..., 3:6].cpu()
# for sp in sPosition:
# s = str(float(sp[..., 0])) + " " + str(float(sp[..., 1])) + " " + str(
# float(sp[..., 2])) + "\n"
# fPone.write(s)
for pp in pPosition:
s = str(float(pp[..., 0])) + " " + str(float(pp[..., 1])) + " " + str(
float(pp[..., 2])) + "\n"
fPone.write(s)
n = 0
pNum = 0
batchsize = 32
count= 0
pts = getSPDensity()
pts = tr.tensor(pts)
iter_per_epoch = int(math.ceil(len(pts) // batchsize))
for i in trange(iter_per_epoch):
# 取空间中的采样点
count = count + batchsize
batchdata = pts[count - batchsize:count]
batchdata = tr.tensor(batchdata)
tr.reshape(batchdata, (-1, 3))
views_chunks = tr.tensor([0, 0, 0]).expand(batchdata.shape).to(device)
tags_chunks = tr.tensor([0, 0, 0]).expand(batchdata.shape).to(device)
raw = run_network(batchdata, views_chunks, tags_chunks, **render_kwargs_train2) # [bs, cks, pts, 3]
alphac = raw[..., 0]
alphac = tr.sigmoid(alphac)
print(alphac)
ptsc = batchdata.cpu()
ptsc = np.array(ptsc)
n = 0
for pt in ptsc:
a = alphac[n]
# si = sn[n]
# print(a)
if a > 0.9:
# # print(pt[0])
pNum += 1
s1 = str(pt[0]) + " " + str(pt[1]) + " " + str(pt[2]) + "\n"
# print(s1)
fPosition.write(s1)
n = n + 1
fPone.close()
fPosition.close()
inds = []
phoneC = o3d.io.read_point_cloud("data/pointCloud/pp.txt", format='xyz')
positionC = o3d.io.read_point_cloud("data/pointCloud/position.txt", format='xyz')
positionS = o3d.io.read_point_cloud("data/pointCloud/sPosition.txt", format='xyz')
phoneC.paint_uniform_color([1, 0, 0])
positionC.paint_uniform_color([0, 0, 1])
positionS.paint_uniform_color([1, 0, 0])
o3d.visualization.draw_geometries([phoneC, positionC, positionS], width=600, height=600)
box_P = []
PCD = []
# ------------------------------
pcd = positionC
pnum = np.asarray(pcd.points).shape[0]
# 加载点云
points = np.asarray(pcd.points)
# 执行K-means聚类
clf = Kmeans(k=5)
labels = clf.predict(points)
# # 可视化聚类结果
# draw_labels_on_model(pcd, labels)
# pcd = pcd.remove_radius_outlier(int(pnum / 30), 0.3)[0]
# with o3d.utility.VerbosityContextManager(o3d.utility.VerbosityLevel.Debug) as cm:
# # -------------------DBSCAN--------------------------
# labels = np.array(pcd.cluster_dbscan(eps=0.2,
# min_points=int(pnum / 5),
# print_progress=False))
max_label = int(labels.max())
print(f"point cloud has {max_label + 1} clusters")
colors = plt.get_cmap("tab20")(labels / (max_label if max_label > 0 else 1))
colors[labels < 0] = 0
pcd.colors = o3d.utility.Vector3dVector(colors[:, :3])
# o3d.visualization.draw_geometries([pcd], window_name="DBSCAN",
# height=480, width=600,
# mesh_show_back_face=0)
n = 0
for i in range(max_label):
ind = np.where(labels == i)[0]
pcd_i = pcd.select_by_index(ind)
pnum = np.asarray(pcd_i.points).shape[0]
# cl, ind = pcd_i.remove_statistical_outlier(nb_neighbors=50*int(pnum),
# std_ratio=0.05)
# pcd_i = pcd_i.select_by_index(ind)
# pcd_i = pcd_i.remove_radius_outlier(int(pnum /5), 0.135)[0]
# if np.asarray(pcd_i.points).shape[0] <= 10:
# continue
n += 1
obb = pcd_i.get_axis_aligned_bounding_box()
obb.color = (1, 0, 0)
box_points = np.array(obb.get_box_points())
box_P.append(box_points)
PCD.append(obb)
PCD.append(pcd_i)
i = 2
# ind = np.where(labels == i)[0]
# pcd_i = pcd.select_by_index(ind)
# pcd_i = pcd_i.remove_radius_outlier(int(pnum / 60), 0.2)[0]
# # if np.asarray(pcd_i.points).shape[0] <= 10:
# # continue
# n += 1
# obb = pcd_i.get_axis_aligned_bounding_box()
# obb.color = (1, 0, 0)
# box_points = np.array(obb.get_box_points())
# box_P.append(box_points)
PCD.append(obb)
PCD.append(pcd_i)
PCD.append(phoneC)
o3d.visualization.draw_geometries(PCD, width=600, height=600)
# KD tree---------------------------------------
# def f_traverse(node, node_info):
# early_stop = False
#
# if isinstance(node, o3d.geometry.OctreeInternalNode):
# if isinstance(node, o3d.geometry.OctreeInternalPointNode):
# n = 0
# for child in node.children:
# if child is not None:
# n += 1
# print(
# "{}{}: Internal node at depth {} has {} children and {} points ({})"
# .format(' ' * node_info.depth,
# node_info.child_index, node_info.depth, n,
# len(node.indices), node_info.origin))
#
# # we only want to process nodes / spatial regions with enough points
# # points = np.asarray(pcd.select_by_index(node.indices).points)
# # if len(node.indices) > 5:
# # SP = tr.tensor(points)
# # tr.reshape(SP, (-1, 3))
# # views_chunks = tr.tensor([0, 0, 0]).expand(SP.shape).to(device)
# # tags_chunks = tr.tensor([0, 0, 0]).expand(SP.shape).to(device)
# # raw = run_network(SP, views_chunks, tags_chunks, **render_kwargs_train2) # [bs, cks, pts, 3]
# # alphac = raw[..., 0]
# # alphac = tr.sigmoid(alphac)
# # ava = sum(alphac) / len(alphac)
# # f = ava < 0.3
# #
# # # early_stop = f
# # if not f:
# # for ind in node.indices:
# # inds.append(ind)
# # else:
# # print(ava)
# elif isinstance(node, o3d.geometry.OctreeLeafNode):
# if isinstance(node, o3d.geometry.OctreePointColorLeafNode):
# # print("{}{}: Leaf node at depth {} has {} points with origin {}".
# # format(' ' * node_info.depth, node_info.child_index,
# # node_info.depth, len(node.indices), node_info.origin))
# points = np.asarray(pcd.select_by_index(node.indices).points)
# if len(node.indices) >= 2:
# SP = tr.tensor(points)
# tr.reshape(SP, (-1, 3))
# print(SP)
# views_chunks = tr.tensor([0, 0, 0]).expand(SP.shape).to(device)
# tags_chunks = tr.tensor([0, 0, 0]).expand(SP.shape).to(device)
# raw = run_network(SP, views_chunks, tags_chunks, **render_kwargs_train2) # [bs, cks, pts, 3]
# alphac = raw[..., 0]
# alphac = tr.sigmoid(alphac)
# print(alphac)
# ava = sum(alphac)/len(alphac)
# f = ava < 0.3
#
# # early_stop = f
# if not f:
# for ind in node.indices:
# inds.append(ind)
# # else:
# # print(ava)
# else:
# raise NotImplementedError('Node type not recognized!')
#
# # early stopping: if True, traversal of children of the current node will be skipped
# return early_stop
#
# pcd = positionC
# point = np.asarray(pcd.points)
# N = point.shape[0]
# # 点云随机着色
# pcd.colors = o3d.utility.Vector3dVector(np.random.uniform(0, 1, size=(N, 3)))
# # 可视化点云
# o3d.visualization.draw_geometries([pcd], window_name="原始点云",
# width=1024, height=768,
# left=50, top=50,
# mesh_show_back_face=False)
# # 创建八叉树， 树深为9
# octree = o3d.geometry.Octree(max_depth=6)
# # 从点云中构建八叉树，适当扩展边界0.001m
# octree.convert_from_point_cloud(pcd, size_expand=0.001)
# octree.traverse(f_traverse)
#
# # pcd = pcd.select_by_index(inds)
# # octree = o3d.geometry.Octree(max_depth=3)
# # # 从点云中构建八叉树，适当扩展边界0.001m
# # octree.convert_from_point_cloud(pcd, size_expand=0.001)
# # 可视化八叉树
# o3d.visualization.draw_geometries([octree], window_name="可视化八叉树",
# width=1024, height=768,
# left=50, top=50,
# mesh_show_back_face=False)
# --------------------------------------------------
# positionC = positionC.remove_radius_outlier(int(pNum / 40), 0.5)[0]
# obb = positionC.get_axis_aligned_bounding_box()
# obb.color = (1, 0, 0)
# box_points = np.array(obb.get_box_points())
# mesh_sphere = o3d.geometry.TriangleMesh.create_sphere(radius=0.25,
# resolution=100)
# phoneC = copy.deepcopy(phoneC).translate((0, 0, -0.25))
# positionC = copy.deepcopy(positionC).translate((0, 0, -0.25))
# positionS = copy.deepcopy(positionS).translate((0, 0, -0.25))
# obb = copy.deepcopy(obb).translate((0, 0, -0.25))
# # o3d.visualization.draw_geometries([phoneC, positionS], width=600, height=600)
# pcd = positionC
# # print(np.asarray(pcd.points).shape[0])
# pnum = np.asarray(pcd.points).shape[0]
# box_P = []
# PCD = []
# with o3d.utility.VerbosityContextManager(o3d.utility.VerbosityLevel.Debug) as cm:
# # -------------------DBSCAN--------------------------
# labels = np.array(pcd.cluster_dbscan(eps=0.2,
# min_points=int(pnum / 100),
# print_progress=False))
# max_label = labels.max()
# print(f"point cloud has {max_label + 1} clusters")
# colors = plt.get_cmap("tab20")(labels / (max_label if max_label > 0 else 1))
# colors[labels < 0] = 0
# pcd.colors = o3d.utility.Vector3dVector(colors[:, :3])
# o3d.visualization.draw_geometries([pcd], window_name="DBSCAN",
# height=480, width=600,
# mesh_show_back_face=0)
# for i in range(max_label + 1):
# ind = np.where(labels == i)[0]
# pcd_i = pcd.select_by_index(ind)
# pcd_i = pcd_i.remove_radius_outlier(int(pnum / 1000), 0.1)[0]
# obb = pcd_i.get_axis_aligned_bounding_box()
# obb.color = (1, 1, 0)
# box_points = np.array(obb.get_box_points())
# box_P.append(box_points)
# print(box_points)
# PCD.append(obb)
# print(len(box_P))
# for i in range(len(box_P)):
# print(i)
# PCD.append(phoneC)
# o3d.visualization.draw_geometries(PCD, window_name="DBSCAN",
# height=480, width=600,
# mesh_show_back_face=0)
08-24
1717
1717
08-25
3135
3135
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
