LiDAR vs Photogrammetrically Generated Point Cloud Data

A 3D mesh created using Global Mapper’s Pixels-to-Points tool displayed in the 2D and 3D views

While both LiDAR and PhoDAR are 3D point cloud formats, the process of creating each is completely different. The nature of the collection process dictates the structural characteristics of the data and its usefulness for specific applications.

In this blog entry, we look at some of the distinct differences between each collection method, and their ideal uses.

A screenshot showing conventional LiDAR data in Global Mapper colorized to represent elevation

LiDAR – The Good

Active Collection Process

Individual 3D points are collected and processed in real time.

More Return Data

Each point includes a range of useful data including return intensity, return count, and classification (added as a post process).

Data Sharing

Data structure has been standardized providing optimal conditions for data sharing and interoperability.

Wide-Area Collection

Scanners mounted on aircraft allow for a wide-geographic area to be captured relatively quickly.

Compact Equipment

Unlike early LiDAR hardware, scanners are now relatively compact and can even be mounted on a UAV.

Ground Detection

LiDAR can penetrate foliage and similar obstructions providing a complete 3D representation of the target area. This allows for ground detection even in heavily forested areas.

Rapidly Evolving Tech

For instance, Geiger-mode LiDAR can provide point densities of 100/sq m or greater.

Accuracy

The points are theoretically more accurate, especially the height value.

DTM Generation

LiDAR is ideal for generating Digital Terrain Models because, unlike photogrammetry, it can “see” through canopies to ground.

LiDAR – The Not So Good

High Cost

Traditional LiDAR requires a manned aircraft to house the necessary hardware.

Sensitivity to Flight Conditions

Collection requires optimal atmospheric conditions for flying. The altitude and speed of the aircraft can also effect the point density.

Poor Anomaly Identification

Raw LiDAR cannot recognize anomalies in the data (e.g. birds underneath the flight path)

Inconsistency in Processing

It is not uncommon to encounter publically available LiDAR files that have been erroneously classified

 

A split image showing a photogrammetrically generated point cloud on the left and a 3D mesh created form this point cloud on the right

PhoDAR – The Good

Minimal Technical Requirements

It’s a more accessible way of creating a point cloud with hardware that can cost as little as $1,000.

On Demand & Versatile Collection

Data can be collected on demand, in a relatively confined area, and with minimal preplanning required.

Greater Point Cloud Density

Point densities are typically much greater than those of traditional LiDAR

Classifiable Data

While not natively LiDAR, a photogrammetric point cloud can have classification values applied and can be exported to a las or laz file.

Raster-Colorized Points

Each point automatically inherits the color from the corresponding images.

DSM Generation

Ideal for Digital Surface Model generation since it is unable to penetrate vegetation like LiDAR can.

PhoDAR – The Not So Good

Requires Distinct Features

Points derived from image analysis require distinct visible features in the geographic area of focus.

Requires Surface Variety

Photogrammetric point cloud generation doesn’t work well when there is a lack of variety in surface texture in images, such as the surface of a desert area or large parking lot.

Requires Sufficient Light

Unlike LiDAR, photogrammetry depends on sufficient ambient light. Clear images are required for generating a point cloud, so shooting images in low-light conditions is not ideal.

Poor Ground Detection

Photographs cannot “see” through canopies like LiDAR can.

Shadows and Sky Don’t Work

Point cloud generation doesn’t work well with images that contain large shadows or a lot of sky.

Accuracy Depends on Ground Control

Horizontal accuracy and elevation values are not as accurate unless ground control points have been used in the processing phase.

Coverage is Usually Limited

Photogrammetric point cloud generation isn’t as practical for large area coverage.

Inconsistent Colors

There is often inconsistent coloring across a surface area due to variations in the color balance of the individual images

More Cleanup

Reflective surfaces can sometimes cause more noise points or anomalies in the data, which would require manual removal. Finer features, such as power lines, may not show up as well as they would in LiDAR data.

Ideal Uses for LiDAR

LiDAR is ideal for collecting data of larger areas and of finer details, such as power lines, pipe lines, and the edges of objects. It’s also ideal for creating digital terrain models, since sensors can penetrate vegetation, allowing for the collection of real ground points.

Ideal Uses of Photogrammetry

Photogrammetry is ideal for surveying smaller areas that contain minimal vegetation. Since it can’t penetrate vegetation like LiDAR, photogrammetry is often better for generating digital surface models, rather than terrain models.

Ideal Software for Both LiDAR and Photogrammetry

Whichever point cloud generation method you choose, Global Mapper and the LiDAR Module are well-equipped to efficiently and effectively process the resulting data. The extensive list of editing, visualization, and analysis tools include point cloud editing and filtering, DTM or DSM creation, feature extraction, contour generation, volume calculation, and much more.

4 Replies to “LiDAR vs Photogrammetrically Generated Point Cloud Data”

  1. I would like to see a webinar that starts from the import of the images from a drone to the export to a Laz file.
    I have a lot of export options out of Drone Deploy but I am finding very hard to great DTM FROM THE DATA.

  2. I am a photogrammetrist (educated and trained, with 20+ years experience) and my office often uses photogrammetric data collected from full scale aircraft and UAVs. We then use this data to create DEMs (bare earth) through photogrammetric point cloud generation and editing. While it is not as accurate as LiDAR under vegetation it is sufficient for many applications.

    Ground control is not necessarily needed if the exterior orientations of the photography are collected to a high level of accuracy. My office gets sub 20cm ground accuracy in open areas with our UAV data and no ground control. The UAV has an accurate GPS\IMU module that collects the EOs of each photo center as the image is taken.

    Photogrammetric collection would not work in a dense rain forest but it collects ground quite well through the vegetation of my region in northern British Colombia Canada. If the ground is visible through the vegetation it can be mapped accurately.

Leave a Reply

Your email address will not be published. Required fields are marked *