Boy asking how to convert Point Cloud data into las laz file

How to convert your point cloud data into .Las / .Laz

published: 11.05.2020

First of all – Don’t worry it’s pretty easy!

Pointly will work with the data format LAS or LAZ. The preferred upload format is LAZ because of its compression, followed by LAS.

Excurse: A LAS (LiDAR Aerial Survey) is an open and binary file format which stores LiDAR point cloud data and a point classification system. It is specified by the American Society for Photogrammetry and Remote Sensing (ASPRS) and its latest Version is LAS 1.4. The purpose of LAS is to function as a standard format between LiDAR tools, including hardware and software. It is seen by the majority as an industry standard for LiDAR data.

Whereas LAS is probably the most used data format for the work with point clouds, there are as well a couple of different formats, like ASCII or PLY. If you have your point cloud data in a format different to LAS, there are some easy ways to convert your specific point cloud format to LAS or LAZ. For example open source solutions like CloudCompare, LASzip or PDAL help you to achieve this within very few clicks.

How to convert benchmark point cloud data sets with Python

Apart from the solutions mentioned above, there is always the possibility to quickly convert any kind of file format using Python and the accompanying LAS packages laspy and Open3D. This approach is very useful, when dealing with custom formats you frequently encounter with benchmark point cloud datasets, such as Semantic3D. In this example, point cloud data is stored in various text files and point classifications are likewise stored separately. To have LAS-formatted files including the classification layer, the “pythonic” way of doing this would be:

1. Read-in the text files using e.g. pandas

raw = pd.read_csv(PATH_TO_DATA_FILE, sep=' ', header=None).values
labels = pd.read_csv(PATH_TO_LABEL_FILE, sep=' ', header=None).values[:,0]

2. Extract the coordinate, the color and intensity information (in this case xyz is stored in the first 3 columns, rgb in columns 5 to 7 and intensity in the 4th column)

xyz = np.ascontiguousarray(raw[:, 0:3], dtype='float32')
rgb = np.ascontiguousarray(raw[:, 4:7], dtype='uint8') * 256
i = np.ascontiguousarray(raw[:, 3], dtype='float32') * 256

3. Create an empty LAS file with the desired LAS-version and format. Define the necessary header attributes offset and scale to fit the data.

hdr = laspy.header.Header(file_version=1.4, point_format=7)
mins = np.floor(np.min(xyz, axis=1))
hdr.offset = mins
hdr.scale = [0.01,0.01,0.01]
outfile = laspy.file.File(PATH_TO_OUTPUT_LAS, mode="w", header=hdr)

4. Write the desired dimensions to the LAS-file – DONE!

outfile.x = xyz[:,0]
outfile.y = xyz[:,1]
outfile.z = xyz[:,2]
outfile.Red = rgb[:,0]
outfile.Green = rgb[:,1]
outfile.Blue = rgb[:,2]
outfile.Intensity = i
outfile.classification = labels
outfile.close()

Now that you know how to convert your data types into LAS, last but not least let’s have a look on how to compress LAS into LAZ. PDAL is one tool for an easy compression. Here (https://pdal.io/workshop/exercises/translation/compression.html) you’ll find a detailed description on how to use PDAL and get that LAZ format.

Download your free infographic

Pointly – Point out what matters.

⇧ back to top