Structured Light Experiments (LiDAR Attempt).

Purpose and Goal:

To build the most capabable LiDAR possible for 50-150 US$.

As I have found the existing solutions absolutely apalling. 400 dollars to buy and dissasemble a Neato vacuum just to get a primative lidar, or spend from 1200 to 6500 US$ for a slow phase differentiating Hykouo lidar, or perhaps 20,000 US$ on a real true time-of-flight lidar.
The need for and lack of a useful lidar for hobby and professional use has now become an intolerable obstacle, if no one else can or will resolve this matter, then i guess ill have too. The problem of robot control and manuevering within human occupied or natural enviroments can only be solved through the use of LiDAR and SLAM. Therefore, without good data (from a LiDAR) it doesnt matter how good your SLAM implementation is. Thus the unavoidable need for cheap, plentiful, and capable LiDAR for robotic fleets of all types (civil and martial).

Solution:
The use of structured light, as demonstranted by the XV-11 lidar in the Neato robotic vacuum. Im hoping for assistance from SparkFun and Parrallax companies and similar minded groups and individuals. I intend for the end device to be open source as much as possible ...

Connection:
im thinking of using SPI for the conection method. So the connection will work as follows :
(Lidar with internal STM32 MCU w/SPI) --> SPI --> (End-User Embedded MCU w/SPI)
Im thinking this will be easier and faster than a crap RS-232 serial TTL 115kbps connection like inb the XV-11.

Previous Threads and Research:

Previous Threads:
<- SLAM for Dummies Like Me!!!
<- Does anyone know how these LiDARs work?
<- CSU Chico Tilt Rotor Flying Machine, (Programming and CPU, 3 of 3).

Links and Research:

Initial Idea:

Specs for Final Device:

Range: 1 foot min, 45 feet max.
Resolution: about 1 inch
Wavelength: NIR or IR, (700nm -1000nm) [Visible (red) for initial calibration only.]
Viewable arc: 360 degrees (i hope)
Enviromental: above freezing, below 150 degree F (*Earth use only), eye safe (i hope) significant unwanted light between 700nm and 1000nm may be a problem.
Size: 4 inch cube
Power: 5v at 300-500mA

Components:

Filter: IR filter passes 700nm and longer, will suppress all visible and UV light.
Main Lens: 12mm diameter, 33mm FL, uncoated, glass.
Main Mirror: TBD, probably a tiny normal mirror cut on a tile saw.
Motor: Probably a DVD/CD motor, about 1200-3000rpm.
Laser: IR 780nm or IR 808nm, laser line, 60 degrees cut down to 20 deg by blinders. 5 and 20mW. (The line will be vertical.)
MCU:Probably STM32 F1.
Detecting Element: Panavision DLIS-2k or DLIS-4k linear pixel senor (monochromatic) [the DLIS-4k is no longer made.]

Cost:

Lasers: 11 and 14 US$
Lens: 4 to 9 US$
filter: 2 US$ (once cut to size)
Mirror: 1 US$ (estimated)
Detector: Unkown
Motor: 2-5 US$ (estimated)
MCU, and PCB: 10 US$ or less
Screws, Case, Wire and Plastic: Aproxiamtley 10 US$\

(I'm trying to keep the build cost to about 50 US$)

Data Sheets and Application Notes:

]app0002_rev_d_appnote.pdf[/file] Panavision DLIS sensor family App Note.

]pds0038revg_datasheet.pdf[/file]Panavision DLIS family Data Sheet. (DLIS-2k)

Technical Points:

A Specific Note On The Filter And Detector:
The filter that i found will pass 700nm and longer, but will block all visible and UV light. The panavision DLIS family is sensitive to a suprisingling limited amount of the IR spectrum. It has been optimised for the detection of viisble light, red specifically, for bar code and fingerprint type machine vision applications. it is insensitive to longer than 1000nm light. Therfore all light that will reach the DLIS senor will be longer than 700nm and shorter than 1000nm, a narrow 300nm passband! i am hoping this will lead to a high signal and low noise condition at the sensor plane.

And for the Lasing Physics:
Within the narrow 300nm passband there will be two possible laser light beams. either a 780nm, 5mw, NIR laser diode module, or a 808nm, 20mW, IR. Both will be line generating lasers. This line will be vertical to the sensor plane, which is horizontal. This arrangement ensures alignment and a narrow but definante fall pattern on the sensor, even with age, vibration and temperature creep.


Supporting graphs for sensor and filter:


Filter, frequency/transmitivity graph.



Frequency sensitivity of Sensor, Panavision DLIS family.


Frome the above graphs we see, the filter will pass 85% of the IR at near 800nm, the DLIS sensor is sensitive to 80% at 780nm and 75% at 808nm compared to red being 100% .

Sat Nov 24 2012, 06:41PM - 1:55AM
First Structured Light Experiments ! (Proof of concept / Technology demonstrator)

(Version 1, Type 1, Mod 10, Gen A) [Success!]



The Setup-up... (A poor college students optical test bench, a plywood table painted white.)



Seen in the above pic, a 12" (y-axis near) target and a 16" (y-axis, Far) Target. Which cuased a visible change of 1.4 inches in the x-axis. The camera was static for the above to pictures, however the Auto Focus did jitter a bit. Still a valid test.



Animated .gif file! Click Me!




Electro-Optical and physical configuration of the would-be sensor.



Far target, a Sharpie paint marker.



Near target, a tube of silver solder.

Lesson learned:
As expected, the width of the laser line as it falls on the DLIS sensor will deteremin the possible vs. actual resolution, as each pixel is
only 4 micrometers wide... (if the laser lights up several pixels, that reduces resolution.)

EDIT: its 1:40am and im drunk, so i might be \wrong, but CAD is indicating i can get the laser width down to 12um at the sensor plane, so that means 3 or 4 pixels will be lit buy my 2mm wide line laser beam... thus about 1 inch of resolution should be possible.

Re: Structured Light Experiments (LiDAR Attempt).
Shrad, Thu Nov 22 2012, 07:40AM

to be eye-safe you have to be around 2000nm or more, if I recall

I would try to source a pulsed laser diode (like those in the same form factor as LEDs) as you can pulse train them and they are cheaper , but I don't know if they exist in this wavelength range

closest affordable model in CW is this one

Re: Structured Light Experiments (LiDAR Attempt).
Shrad, Fri Nov 23 2012, 12:10PM

there are risks, and it's controlled by the CDRH

your device would be considered as a laser scanning device and there is a MPE (max permissible exposure) and other fail-safe concerns

read this carefully :

and get other resources from their website about security

IR starts being eye-safe around 1500~2000nm as it's being absorbed by the eye surface rather than the retina (transparency of the eye cristaline)

Re: Structured Light Experiments (LiDAR Attempt).
Patrick, Mon Dec 10 2012, 08:38PM

Very cool work !!!


iliasam, ive PM'ed you with some details.

I would like to work with others on this matter, if possible.
Next, how did you find out about my work?
and finally, your work is cool, it appears you have considered how we are to get a SLAM model working. should we use desktop/laptops, or embedded systems?

I see even with only 128 pixels, your still getting good data. youll have to see my thread here : (there are two pages)
which documents the "range problem" and yes more pixels with a smaller pitch helps...

A few questions if you dont mind:

-how did you align your optics?
-are you using a line or dot laser? (one of your pics in the dark shows dots)
-are you using the STM32 MCU ?
-whats your RPM on your belt driven turning table ?
-How are you getting the data to the base robot? I plan to use IR .
-which pixel sensor are you using?
-what wavelength of light?

that tricylce bot is cool too!

Re: Structured Light Experiments (LiDAR Attempt).
Patrick, Wed Dec 12 2012, 05:28AM

New Work !

first visible light test arragement with red laser line generator.

12.7mm dia lens, with FL of 33mm. red laser light and its epoxied to a 1.5mm acrylic plate.

(the final version for IR will be epoxied to the 1.5mm thick filter.)

I shot the laser light in directly, so im not sure the beam width is accurate. beam length for the above test is 134 inches long, and in the pic its 0.140 wide...[thus 3.56mm wide @ 3.4m] and the thin beam is 0.033 " (or 0.8mm) wide.


EDIT: the O'great CAD gods are telling me that at 2.9m a 1.5mm beam will be 18um wide at the sensor plane, which is about 4 pixels wide.

Re: Structured Light Experiments (LiDAR Attempt).
Shrad, Mon Dec 17 2012, 01:52PM

Patrick,

While searchink for laser diode driver papers I stumbled upon this :

I think it is really close to what you are attempting and contacting them would raise a flag on several issues you could face, so more energy is spent polishing the design instead of bumping on obstacles

Re: Structured Light Experiments (LiDAR Attempt).
Patrick, Mon Jan 14 2013, 01:32AM

ok new work completed...

ive got two line generating lasers, one will be kep as a backup, in case i kill the other.

for 12 US$ i have a small barrel sized IR laser. I cut the case open to learn its made out of nickel-plated copper, which explians why its too dam heavy. having hack-sawed the case open ive found the back 80% of the tube filled with a dense populated circuit board, so thats a relief. the driver is fully integrated and working.

i wish i could figure out a way to rotate the beam and view port with out rotationg the sensor and laser, since moving brush contacts strike me as unreliable.










EDIT: i have removed half the wheight by means of cutting off the copper, but a thought occurs : that copper was the heatsink! DAM!

Re: Structured Light Experiments (LiDAR Attempt).
iliasam, Wed Jan 16 2013, 01:56PM

i wish i could figure out a way to rotate the beam and view port with out rotationg the sensor and laser, since moving brush contacts strike me as unreliable.

As far as I know, SICK lidar have a special scanning mirror:

But I think that this scheme is not suitable for triangulation lidar.
Hokuyo lidar seems to have brush contacts: http://www.iheartrobotics.com/2010/03/under-hood-secrets-of-hokuyo-urg-04lx.html
Heato lidar have brush contacts.
Also i'm thinking that only brush contacts can give 360 scanning angle.
I have found this contact brush, it is looking interesting: http://www.adafruit.com/products/736

that copper was the heatsink!

I think that if your laser's power less then 5mw, laser and it's driver does not need heatsink.

Re: Structured Light Experiments (LiDAR Attempt).
Carbon_Rod, Thu Jan 17 2013, 08:56AM

...still standing firm like the Oak?


Have you considered blinking the laser off, and doing image background subtraction to recover your signal line?

Best of luck,
Rod

Re: Structured Light Experiments (LiDAR Attempt).
Patrick, Thu Jan 17 2013, 09:21PM

Carbon_Rod wrote ...

...still standing firm like the Oak?


Have you considered blinking the laser off, and doing image background subtraction to recover your signal line?

Best of luck,
Rod

yes i am already implementing the blink rate into my design, and the DLIS-2k is compatible with image subtraction at high speed.

Re: Structured Light Experiments (LiDAR Attempt).
iliasam, Mon May 27 2013, 10:27AM

Have you seen this project: ?
They have already done a prototype with good characteristics and want to start production.
The only question is the price - they did not give it.

Re: Structured Light Experiments (LiDAR Attempt).
Linas, Fri Jun 07 2013, 05:24PM

hey, does any one know fast linear photo-diode sensor ?
128 or 65 points, with 8-20MHz readout CLK.
for now i use TSL1401, and i get 58KHz read speed with FFT, but i need to be able to get 100-200kHz cycle speed.
Best what i was able to find is this:


But i can't get, can't buy, or get samples
and fastest for 128 pixels would be CCD , but it will be expensive to drive (ultrafast multiplexing opamp for single adc readout, and CMOS to MOS logic level convertion is tricky (3.3 to 12V at 10MHz)

Re: Structured Light Experiments (LiDAR Attempt).
Linas, Sun Jun 09 2013, 01:05PM

iliasam wrote ...

What about iC-LF1401? The maximum pixel readout frequency is 5MHz. It's 2.5 times faster than TSL1401.

Hm, i don't think so
TSL1401CL:



Oh, i get myself again, as it turns out, i can't find any RL0256EEQ-011 Datasheet ( I had for D, not E series),
Any idea how to get it ? i try to contact PerkinElmer but with no result ...

Re: Structured Light Experiments (LiDAR Attempt).
Patrick, Sun Jun 16 2013, 04:43AM

a well said statement (as pertaining to the XV-11 LiDAR ):
In response to a question, as to what LiDAR really "is"...


[Source:

by chrlsrchrdsn - 12/10/2010 - 11:36pm



Either ToF or Triangulation are LIDAR. The concept is to shoot a specific beam of light from a laser (though technically you don't have use a laser, you could use something like a pinhole focused LED, but the range would be very short) then use the return from that specific beam to generate the results. Time of Flight is most commonly used until now because high speed CMOS sensors were not available in the 1990s. However precise clocks and a sensor that could tell it got hit fast enough was. After a bit, instead of using a clock you could actually read the charge level of the recharging laser pulse capacitor at the time the sensor said it was hit and give a distance reading since moving electrons and light aren't too far off in speed.


BUT now CMOS sensor of sufficient speed and with global shutter make triangulation possible. (Global shutter is a way to "gate" the picture by recording all pixels simultaneously. No having to read in the lines to get the results that could be changing as you read.) Triangulation is less sensitive to beam scatter than ToF, and triangulation can actually be more accurate for sub-mm measurement.


Okay, the simple answer, triangulation is a recognized medthod for LIght Detection And Ranging.

Login or register to post comments