Public Lab Research note


Options for Low-Cost LIDAR

by ajawitz | September 28, 2014 15:58 | 9,049 views | 9 comments | #11196 | 9,049 views | 9 comments | #11196 28 Sep 15:58

Read more: unstable.publiclab.org/n/11196


What I want to do

I want to research the current availability of low-cost LIDAR sensors and their potential integration into existing citizen science/grassroots mapping efforts.

Why I'm interested

The use of laser rangefinding devices has produced nothing short of a revolution in fields ranging from archeology to land use and geology. For example, high-end expeditions in places like Angkor Wat and Guatemala have used LIDAR to uncover ancient structures that have been hidden beneath the rainforest for thousands of years. Yet, like many other research methods, only well-funded institutions have been able to afford the underlying technology. It is precisely such characteristics that make this technology ripe for disruption!

My attempt and results

A quick search through past PLOTS discussions reveals that such a capability is indeed highly desirable across the spectrum of citizen science, but until recently it seems that the phrase "affordable LIDAR" was somewhat relative. For example a "hobby LIDAR" has been available for roughly $400, and this may indeed make the technology accessible to far more organizations than would have previously, but my definition of "accessible technology" would require a price range of under $100.

Sure enough, it does indeed appear that such a device called LIDAR-Lite successfully met its crowdfunding goal and will soon be available for $72 at https://store.3drobotics.com/products/lidar-lite/.

According to a post on Hackaday

it can sense objects out to 40 meters with 5% 95% accuracy, communicates to any microcontroller over an I2C bus, and is small enough to fit inside any project.

Questions and next steps

I am interested in hearing if others have looked into such a capability, if other options are available, and/or how it might be integrated into existing projects. The fact that the LIDAR-Lite communicates over the i2C bus means it should be feasible to incorporate into my design for a low-cost mapping UAV alongside the camera. As such a vehicle can be programmed to fly a plotted flight plan, it potentially allows for more precise calibration between LIDAR, Infrared and standard imagery.

LIDAR can also be incorporated into non-powered mapping platforms such as balloons or kites though it would entail the addition of a MiniPC or microcontroller and the additional weight required by batterries etc...

UPDATE-6/29/15

The LiDAR-LITE is now widely available for less than $100 through many different outlets. I should be receiving one by tomorrow 6/30/15! Since this note was originally published, I have been developing a quadrotor to serve as a test bed for various sensors and components that will eventually be incorporated into my build for a low-cost flying wing UAV. The quadcopter was far easier to build as the platform's popularity produces more readily available components and documentation. Nevertheless, it has taken me almost a full year to get to the point I am at now. At present, the quad is equipped with a full GPS navigation system running on a Flip32+ Flight Controller (incl. Magnetometer and Digital Compass). A 3D Printed camera gimbal can be automatically set to keep the Mobius Action Cam positioned downward.
Quadradome.jpg

All that remains before it should be ready for first flight is a new RC Transmitter as my Dx5e is too limited in its controls. After I achieve first flight, I intend to create a detailed research note with step by step instructions aimed at non-experts so they can hopefully avoid a full year of trial and error of their own!

Platforms I will be testing include the LiDAR-LITE with a Raspberry Pi A+, Open Source transmitter firmware including OpenTX and OpenLRS as well as various aerial mapping configurations with the Mobius ActionCam.

Fingers Crossed that the first flight isnt also the quad's first crash!


9 Comments

Aloha, I recently started using an UAV to capture high resolution imagery that I then process with Pix4D software. One of the outputs of the software (provided you have good image overlap) is a .LAS (or .LAZ) file that has an resolution about 4x the 2D pixel resolution. So if my image resolution is ~3.5 cm, then my .LAS file pixels have a ~6 Inch pixel size. No need for a LIDAR sensor, all done via Photogrammetry.

Reply to this comment...


There have been a few discussions about whether or not one could use OpenCV for 3D Orthomosaic mapping, but beyond that I've yet to find any open source solutions for Photogrammetry or 3D Orthomosaic. Nor have I discovered any turnkey proprietary solutions that cost under $70. Nevertheless, its good to know that a software approach is possible. Even if the notoriously complex OpenCV is the closest thing we have to an open source option.

Reply to this comment...


I think MicMac will take LIDAR data http://publiclab.org/notes/mathew/02-25-2014/3d-data-from-image-sets-autostitching

Reply to this comment...


Surely the most valuable point of Lidar is to provide high resolution data of areas that are inaccessible by other methods. Such as dense vegetation and tree cover. The major issue will be processing power. This is usually provided a stand alone system rather than a flight controller. The amount of data collected is very large, but as usual technology will improve as time moves on and i, like the rest, will be waiting in the wings.

Reply to this comment...


UPDATE:6/29/15 The Lidar LITE can now be purchased for under $100 through Sparkfun, Amazon etc... I just recently found one at Robotshop.com on sale for $75 and it should be arriving tomorrow! I intend to use it on my Naze32 quadcopter that I've been building for almost a year. It is finally in its final stages and if all goes well, I'm hoping to have it in the air by next week!

Quadradome.jpg

While I intend to post detailed documentation on the build once it gets off the ground, I wanted to start looking into ways to incorporate the LiDAR LITE into aerial mapping for which I have been using a Mobius ActionCam. @smathermather- Does #OpenDroneMap support the LiDAR-LITE?

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


Great project - I'm interested in seeing what the results of the LED lidar would be. I think it probably needs a laser to be useful but maybe with good optics it doesn't? Let us know!

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


I do believe @pinpointmapping has the right idea. Unfortunately Pix4D software is orders of magnitude more expensive than $100 even if you go with the free crippled version because of the processing hardware requirements. Perhaps there is some open-source package available that works as well but I have not found it yet. Perhaps the http://www.adapteva.com/parallella/ will get some software written for it that will do what we want.

The "LIDAR" you are referencing is probably better labeled as a laser/LED altitude sensor. The problem you will have with it is that the field of view will be around 20 degrees, as one of the devices you reference states. That will result in a large pixel size which, of course, is dependent on the altitude. Get a protractor out and see what area is included with different distances. A real LIDAR will have a laser beam (think laser pointer) scanned over the viewing area and each spot that the laser shines on will be used to measure height resulting in a much smaller pixel size. (This is a little simplified but broadly accurate.)

Any way I am interested in your results we may all learn something from your endeavors.

Reply to this comment...


It looks like somebody finally did do some point cloud mapping work with the LidarLITE- http://myrobotlab.org/content/lidar-lite-distance-measurement-and-creating-point-cloud The results do seem to confirm @danbeavers hypothesis about resolution. Either way, it gives me enough to work with and start doing some of my own experiments. Even if it isnt the "Peoples LiDAR" dream come true, it still represents a pretty big leap over most rangefinders in current use. So there's a lot of territory to cover for sure!

Reply to this comment...


Hi, Really interesting work. Any new updates on this project? I'm wondering if there are any DIY UAV / LiDAR options that would be accurate enough to identify and map possible cultural heritage sites.

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


Login to comment.