Did you forget to `#include ＜pybind11/stl.h＞`? Or ＜pybind11/complex.h＞,＜pybind11/functional.h＞

Posted 2021-10-09 沉迷单车的追风少年

项目场景：

基于深度学习的三维点云可视化

---

问题描述：

Traceback (most recent call last):
  File ".\\draw_npy.py", line 25, in <module>
    o3d.visualization.draw_geometries([pcd.points])
TypeError: draw_geometries(): incompatible function arguments. The following argument types are supported:
    1. (geometry_list: List[open3d.cpu.pybind.geometry.Geometry], window_name: str = 'Open3D', width: int = 1920, height: int = 1080, left: int = 50, top: int = 50, point_show_normal: bool = False, mesh_show_wireframe: bool = False, mesh_show_back_
face: bool = False) -> None
    2. (geometry_list: List[open3d.cpu.pybind.geometry.Geometry], window_name: str = 'Open3D', width: int = 1920, height: int = 1080, left: int = 50, top: int = 50, point_show_normal: bool = False, mesh_show_wireframe: bool = False, mesh_show_back_
face: bool = False, lookat: numpy.ndarray[float64[3, 1]], up: numpy.ndarray[float64[3, 1]], front: numpy.ndarray[float64[3, 1]], zoom: float) -> None

Invoked with: [std::vector<Eigen::Vector3d> with 2967 elements.
Use numpy.asarray() to access data.]

Did you forget to `#include <pybind11/stl.h>`? Or <pybind11/complex.h>,
<pybind11/functional.h>, <pybind11/chrono.h>, etc. Some automatic
conversions are optional and require extra headers to be included
when compiling your pybind11 module.

---

原因分析：

这个error提示看得我一脸懵逼，在反复折腾之后，终于搞定。

核心是这句话：

o3d.visualization.draw_geometries([pcd.points])

改成

o3d.visualization.draw_geometries([pcd])

---

解决方案：

原先报错的代码：

import numpy as np
import open3d as o3d

filePath = './GEN_Ours_chair_1631677128/out.npy'
# filePath = './GEN_Ours_airplane_1631630214/out.npy'
np.set_printoptions(suppress=True) # 取消默认科学计数法，open3d无法读取科学计数法表示

loadData = np.load(filePath)
print("----type----")
print(type(loadData))
print("----shape----")
print(loadData.shape)
print("----data----")
print(loadData)

# source_data = np.load(filePath)[:,0:3][:,0:3]
point_cloud = o3d.geometry.PointCloud()
source_data = np.load(filePath)
# source_data = np.squeeze(source_data)
print(source_data.shape)
# print(source_data)
pcd = o3d.geometry.PointCloud()
source_data = source_data[:, 0:3].reshape(-1, 3)
pcd.points = o3d.utility.Vector3dVector(source_data)    # numpy.array转换成point cloud
o3d.visualization.draw_geometries([pcd.points])

print("finish!")

改正后：

import numpy as np
import open3d as o3d

filePath = './GEN_Ours_chair_1631677128/out.npy'
# filePath = './GEN_Ours_airplane_1631630214/out.npy'
np.set_printoptions(suppress=True) # 取消默认科学计数法，open3d无法读取科学计数法表示

loadData = np.load(filePath)
print("----type----")
print(type(loadData))
print("----shape----")
print(loadData.shape)
print("----data----")
print(loadData)

# source_data = np.load(filePath)[:,0:3][:,0:3]
point_cloud = o3d.geometry.PointCloud()
source_data = np.load(filePath)
# source_data = np.squeeze(source_data)
print(source_data.shape)
# print(source_data)
pcd = o3d.geometry.PointCloud()
source_data = source_data[:, 0:3].reshape(-1, 3)
pcd.points = o3d.utility.Vector3dVector(source_data)    # numpy.array转换成point cloud
o3d.visualization.draw_geometries([pcd])

print("finish!")