Welcome to the MATT Project home page. We are going to go out on a limb and trying something very different. Instead of reporting on a successful experiment in the development of a new tool we will be posting the day-to-day progress with the MATT Project. Our trimphumps, our failures and hopefully our successfully completed device for making RTIs/PTMs of artifacts in the field or artifacts too large to fit on PATT in our lab. The following will be a progress report and hopefully take you along on an interesting adventure! The MATT Project Journal This is were you can watch the development of MATT, the Mobile Artifact Turn Table to be used in the creation of RTIs/PTMs. We have listed this as an ArchX Project because of the learning possibilities. Once completed we will post an ArchTools page. A Word About RTI/PTM Reflectance Transformation Imaging is the general term for photographic imaging techniques that take advantage of racking angle light. PTM stands for Polynomial Texture Mapping which is a mathematical process developed by HP Labs that greatly enhances the contrast achieved with racking angle light. Part 1: Getting Started The first part in the development of a project is to set some parameters. We will be somewhat loose about these as to allow for changes during development. Jan 1, 2015 We begin by laying two meter sticks on the floor perpendicular to each other in order to decide a practical artifact size. I next began experimenting with my SLR with different lenses and various shooting distances. After several goes at it I decided to begin the MATT Project with my Sigma 10-20mm wide angle lens at a shooting height of 60cm which created a 45cm x 75cm viewable area. Jan 2, 2015 I made a sketch of what MATT in action might look like. I like to begin projects using Visio. With Visio I can quickly sketch and annotate a new concept. Jan 4, 2015 Once we have an idea of what MATT will be doing we need to create a map by which we expect to proceed. Once again this will change as we progress. Here, again, I find Visio exceptional at creating flow charts. Jan 13, 2015 Back from SHA in Seattle! Had a great time and made many new friends! Time to get back to work. The following diagram is our first stab and a functional flow chart for MATT. This is a visual on how we are thinking MATT should function. The yellow circles are operator driven decisions. The green boxes represent algorithms yet to be written. We must digest this a little then begin work on the algorithms. Jan 15, 2015 First Miss-Step Last night while reading a paper sent to me by Dr. Harold Mytum, Professor of Archaeology, University of Liverpool, I realized I may have made a design error with MATT. Dr. Mytum and I are collaborating on raising the awareness of archaeologists to the benefits Reflectance Transformation Imaging in the field and the laboratory. The Error As you can see in the first sketch of MATT, I have the camera and the Arm Rotation Device mounted to the same support. In “Multi-light Imaging for Heritage Applications” which Dr Mytum set to me stresses the importance of the camera remaining completely stable during the image taking portion of the process. I fear that vibrations from the Arm Rotation Device will cause slight movement of the camera and result in poor pixel alignment in the final image. PATT is designed to allow different light sources to be swapped in and out. This is one of the advantages of PATT, with only one source of light there is only one lamp to exchange. MATT will also use only one source of light, however, after looking over another paper that Dr. Mytum sent, I can see the advantages of being able to attach either a bright LED light source or a remote flash unit (for daytime RTI work.) Also having an array of neural density filters will be necessary. You use the ND filters to darken the artifact in daylight and then the flash will create a bright racking light source for the RTI. The Fix We shall have to design the Arm Rotation Device to mount on a separate tripod, behind the camera. What Can We Take Away From This There are several lessons to be learned from our first miss-step.
- The value of collaboration: With input from others you are more likely to gain their insights early in the design process and make design changes before much physical work goes into a project.
- The value of networking: I met Dr. Mytum at this year’s Society of Historical Archaeologists, SHA, annual meeting (early January). I had several stimulating conversation with Dr. Mytum and many other archaeologist. It is important to attend good conventions and to get out there and talk to folks.
- Business Cards: I cannot stress enough how important it is to carry and pass out your business cards at conferences. On my business card I state my affiliation with Wayne State University, clearly stating who I am and what my specially/interest is: RTI/PTM Specialist, contact information, websites authored, and on the reverse I have before-and-after screenshots of an artifact and the resulting RTI. This makes it very clear the type of research I’m involved in. And don’t be afraid to jazz up your card. You get one shot at an introduction – give them something to remember you by.
- Follow Up: Follow up on the contacts you make with a simple email. State the nature of the conversation you had and how wonderful it was to meet the individual.
- Network Through the Internet: When you are designing a project you might be surprised at how willing people are to assist you in your research. When I began working with RTI/PTM I wrote an email to Dan Gleb of HP Labs. He was most helpful and encouraging and gave me resource suggestions. Other professors and researchers have also been helpful. I must also stress the helpfulness of the “Maker Community.” If you are not familiar with the “Maker Community” I suggest you read about the OpenROV (www.openrov.org) project and learn about the open sources of information available on the web.
- Blogging: Good research should be shared and blogs can be a great way to do this. You are here because I have a blog…enough said.
Jan 30, 2015
Several interruptions have caused delays in the MATT Project. I plan on using my 3D printer to fabricate some of the parts for MATT, and it took a nose dive. It turn out that two, not just one, but two USB cables were defective causing me to look else ware for the communication problems with the 3D Printer. Three mother boards later I discovered this most unlikely scenario. The other difficulty was an un-backed up hard drive that crashed.
I had been doing some preliminary code work for the microprocessor for MATT when my laptop crashed. Luckily my colleague, Mark Jazayeri, suggested a fix that worked.
Hopefully, we will now be back on track with our research.
Feb 5, 2015
Following is the new diagram for MATT.
The change is simple but critical. We have isolated the camera from the rotating light arm. This will prevent vibrations caused by the movement of the light arm from causing camera movement which would result in pixel mismatches is the final PTM.
From our new diagram we can begin to compute some of the critical dimensions. We will allow a generous 20cm for the camera and lens, 4cm of free space between the Rotation Motor and the Camera, and 60cm from the surface of the artifact to the lens. With a 70cm radius for the artifact this results in approximately a 1m light arm.
Arms & Motors
Once we have computed the length of the arms, we need to pick out the arm material and the motors to move the arms.
I have decided to go with 12 inch 8020 extruded aluminum stock. It is fairly light weight and strong.
8020 12 inch Extrusion
Special T-nuts slide into the grooves and screws can then be attached to the T-nuts.
For the motors I have chosen three types, the HiTEC SPG805A and the HiTEC HS-755MG motors both with 180o rotation, and the VEX 393 motor for the turntable. These are fast servos with lots of torque. The following diagram shows the motors, turntable and arm positions.
Let’s do some simple math:
Torque is the rotational force at a given arm length. The relationship between force and length of arm is an inverse one. Think of how levers work. The longer the distance between the fulcrum and the load the more force that is required from the input end of the lever given that input length has not changed.
When we are talking about torque we must think of the fulcrum point as being the shaft of a motor. This shaft exerts rotational force called torque. This essentially means that the input arm of our lever system is very short. If our motor is rated at 200gm.cm it would be able to lift 200gm on an output arm length of 1cm. If we increase the output arm length to 20cm then we must divide the torque by 20 which would equal 10gm of lifting force.
Torque: 1751in.oz = 121970gm.cm
=1219.70gm of force on a 100cm arm
This motor will be used to lift the entire arm and lamp assembly. We will be leaving the arms slightly longer than needed in case we have to use a counter-weight to offset the weight of the entire arm assembly. (See the Motor/Arm Diagram above.)
HS-755MG Servo Motor
Torque: 14.4kg.cm = 14.4kg of force at the end of a 1cm arm.
= 14400gm of force at the end of a 1cm arm
= 144gm of force at the end of a 100cm arm
The first of these motors will be used to lift the second half of the arm and lamp and a second one will be used to rotate the lamp to keep it pointed at the artifact.
Vex 393 Motor with a Turntable
This is the same motor-turntable combination used in PATT. It is quite powerful and I think it will do the job.
These four motors should be enough to move a lamp or flash assemble around the virtual dome that MATT will create.
The 8010 aluminum extrusions, the two HS-755MG motors, and the 393 Motor and Turntable have arrived. We are expecting the SPG-800A motor in the next few days. We will begin designing the motor mounts next.
Latest: MATT is still under construction. I had a snafu with the motors and arms that I think I have rectified with lighter arms and stronger motors.
The main delay, however, was due to a good thing. We got our 3D printer up and running and are now able to print parts for MATT! Pictures will be coming soon. Also I am printing a new and larger printer which will help with the larger parts needed for MATT.
The following diagram shows our progress to date.
Site Under Construction: More to come!