Point cloud processing in Computer Vision - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to load a point cloud file using Open3D.

Computer Vision

import open3d as o3d

pcd = o3d.io.read_point_cloud([1])
print(pcd)

Drag options to blanks, or click blank then click option'

Apointcloud

B"pointcloud.ply"

Cread_point_cloud

Do3d

Attempts:

3 left

💡 Hint

Common Mistakes

Forgetting to put the filename in quotes.

Passing the function name instead of a filename.

The read_point_cloud function requires the filename as a string to load the point cloud data.

2fill in blank

medium

Complete the code to downsample a point cloud using voxel grid filtering.

Computer Vision

downsampled_pcd = pcd.voxel_down_sample([1])
print(downsampled_pcd)

Drag options to blanks, or click blank then click option'

Avoxel_size

B5

C0.05

D"0.05"

Attempts:

3 left

💡 Hint

Common Mistakes

Passing the voxel size as a string instead of a float.

Using an integer instead of a float.

The voxel_down_sample method expects a float representing the voxel size, such as 0.05.

3fill in blank

hard

Fix the error in the code to estimate normals for the point cloud.

Computer Vision

pcd.estimate_normals(search_param=o3d.geometry.KDTreeSearchParam[1](radius=0.1, max_nn=30))

Drag options to blanks, or click blank then click option'

ARadius

Bradius

Cradius_size

DRadiusSearch

Attempts:

3 left

💡 Hint

Common Mistakes

Using lowercase 'r' in 'radius' causing attribute error.

Using a wrong class name like 'RadiusSearch'.

The correct class name is KDTreeSearchParamRadius with uppercase 'R' in 'Radius'.

4fill in blank

hard

Fill both blanks to create a point cloud from numpy array and visualize it.

Computer Vision

import numpy as np
import open3d as o3d

points = np.random.rand(100, 3)
pcd = o3d.geometry.PointCloud()
pcd.points = o3d.utility.Vector3dVector([1])
o3d.visualization.draw_geometries([[2]])

Drag options to blanks, or click blank then click option'

Apoints

Bpcd

Cnp.array

Dpoints.T

Attempts:

3 left

💡 Hint

Common Mistakes

Passing raw numpy array directly without wrapping.

Passing points instead of pcd to draw_geometries.

The points attribute expects a Vector3dVector wrapping the numpy points array, and draw_geometries needs the point cloud object.

5fill in blank

hard

Fill all three blanks to compute and print the axis-aligned bounding box of a point cloud.

Computer Vision

aabb = pcd.[1]()
min_bound = aabb.[2]
max_bound = aabb.[3]
print(f"Min bound: {min_bound}, Max bound: {max_bound}")

Drag options to blanks, or click blank then click option'

Aget_axis_aligned_bounding_box

Bmin_bound

Cmax_bound

Dget_bounding_box

Attempts:

3 left

💡 Hint

Common Mistakes

Using a wrong method name like get_bounding_box.

Confusing min_bound and max_bound attribute names.

The method to get the bounding box is get_axis_aligned_bounding_box, and the bounding box has min_bound and max_bound attributes.