P19510: Eastman Museum Digitization Process

Subsystem Build & Test

Table of Contents

Team Vision for Subsystem Level Build & Test Phase


Subsystem 1: Table with rollers

To create the table we utilized table tops donated by the Eastman Museum. Harley and Rahul purchased 2 X 4’s from home depot to build the legs to support these table tops. To create the legs a sort of corner approach was used. By ripp cutting all of the 2 X 4’s at a 45 degree angle down the length of the leg we could then wood glue and screw together 2 ‘legs’ to create a 90 degree angle leg. This leg design was used to provide maximum support to the table and the ensure the table wouldn’t wobble back and forth while the conveyor is running. 2 sets of 4 legs were attached to one another using 4 more 2 X 4’s running along all 4 sides of the table to once again reduce wobbling and create a base for attaching the table tops using a 90 degree angle bracket on the bottom of the table top. Once the two tables were built and a few beginning pockets were sawed in to allow the range finders to be placed we began looking at the rollers. While the initial plan was to adopt a wooden roller designed by Chris, after further consideration, it was found to be too heavy and a little too robust for the application. So Harley and Rahul went about purchasing and modifying a piece of 2 inch PVC pipe to create new rollers. These rollers were created by sawing a PVC pipe in to 2 pieces of 14.5 inches and then drilling 4 holes to place the ‘studs’ for the eyelets. 2 holes were also drilled into the flat ends of the PVC pipes (the caps) so that a ⅜ inch -16 TPI threaded rod could be run through the PVC pipe. Once the rollers were designed, we went about creating the brackets to support the rollers. Yet again, the original design was for the 4 inch diameter solid wood rollers which was too bulky, compared to the now 2 inch PVC pipe rollers. So, a long piece of an L bar was found downstairs and Harley and Rahul determined that this could be modified to act as the roller holders. Using this L bar, individual brackets were cut and holes were drilled through the brackets to allow for mounting on to the table legs, table top, and to each other in certain cases. Once all of these brackets were machined, they were connected to the table and the rollers were mounted.

Table with Legs

Table with Legs

Table with Rollers

Table with Rollers

Subsystem 2: Conveyor Belt

To create the conveyor belt itself a 4.5 yard piece of black upholstery was cut in to strips of 9 inches to be hemmed along the sides and sewn together. The conveyor belts were cut by Harley and Rahul using a rig set up to hold a wood stop at 9 inches away from a gap where the scissors were placed, the fabric was then slowly moved along the table and wood stop and through the scissors to create as close to 9 inch strips as possible. Once the 6 strips were cut Erin and Harley went about pinning and sewing a 1.25 inch hem on both sides to create a 5.5 inch conveyor belt in the end. After both sides of the fabric were hemmed two strips were connected end to end to one another to create the full conveyor belt. The original design of the conveyor belts relied on eyelets which would be place every 4 inches along the sides of the belt and would then be hooked along the roller using the studs placed through the PVC pipes. However, through testing and observation we noticed that the elastic forces in the belt material were high enough to hold the conveyor and to create enough friction to cause the belt to move as the roller turned so that no studs and eyelets were needed and these additions were removed. After attaching one side of the conveyors we also determined that it would be important for the other side to be able to come apart if needed so that we can run the entire conveyor belt around Elizabeth’s current copy stand without having to cut the tabletop off first. Our original idea was to use a zipper but then Erin and Harley came across some hooks and loops so that the fabric could sort of hook together. This creates a less noticeable attachment point and less disruption on the rollers. FInally, the conveyor is driven by a DC motor mounted at the end of the discharge table. The rollers are fixed to a threaded rod using a pin, and that threaded rod is what is used to drive the subsystem. Chris machined a special adapter which allows the threaded rod and motor shaft to connect. The threaded rod is simply threaded into one end of the adapter, while the motor is locked in place with a set bolt on the other to avoid slipping.

Cutting Conveyor Belt: Cutting Fabric

Conveyor Belt Measurements

Conveyor Belt Measurements

Conveyor Motor Driving Mechanism

Conveyor Motor Driving Mechanism

Subsystem 3: Camera Height Adjustment

The camera height adjustment is a modification of the existing column on the back of the photographing table. Like the conveyor, the height adjustment is powered by a motor mounted to the top of the camera-column mount. The handle that is currently used to adjust the height will be removed, and a sprocket will be placed on the end of the shaft. We decided that we will machine an adapter that allows the sprocket to easily attach to the end of the adjustment shaft. The idea is that it will be removable if there proves a need. This same premise is used for the motor shaft → sprocket connection. The adapter is securely mounted to the flat part of the motor shaft and the sprocket using a pins to avoid slipping. The motor is supported using a custom machined 90 degree “clamp” which holds the motor in place and is attached directly to the top of the support surface. Using an allen wrench, the user can loosen the two halves of the motor support for maintenance and repair as seen fit. Additionally, the knob that will be removed is tentatively planned to be placed on the other side of the height adjustment shaft, should Elizabeth decide she wants to change the height of the camera manually. For this subsystem, all custom parts were designed & machined by Chris.

Motor Height Adjustment Annotation

Motor Height Adjustment Annotation

Subsystem 4: Light Stands

To create the light stands we used a 4x4in by 48in long piece of lumber as the base, 80/20 T-slotted aluminum framing and 2x4’s for the legs and arm to make a C-shaped stand that can go around the table. Rahul cut the piece of lumber in half and added metal fins to the bottom of the base to prevent the stand from tipping over. To create the leg, we used a combination of 80/20 and 2x4’s, the top half is 80x20 to allow for adjustments in the height of the light. The 80/20 was cut to 23in and the 2x4’s was cut to 32in. To hold up the 80/20 bar we sandwiched it in between 2 pieces of wood and bolted it through the center, as an extra precaution we added a bracket to the back to hold everything together. Next, we screwed the leg to the base by cutting out a 1x3.5in by 2 inch deep section on both sides of 4 inch side so the leg rests in place while still maintaining the grip at the 80/20 connection. Next, the 80/20 arm that holds the light over the table was attached with a bracket and a diagonal brace that Rahul cut from 80/20. To attach the lights to the arm Rahul and Tyler talked to Bob Rose, an assistant professor in the photo science school, about attachments that are commonly used in current light stands. After talking to him we were able to look for adapters that fit the current lights. Luckily the George Eastman museum had spares adapters that can easily be placed in the 80/20 bar.
Light Stands

Light Stands

Subsystem 5: Image sizing

Sizing of the trays and artifact with the webcam is relatively robust in low-light conditions.
Lighting test setup

Lighting test setup

The trays, however, needed to be lined with white on the edges that intersect with the conveyor.

Lined Tray

Lined Tray

More improvement is needed on the tray sizing program to improve its robustness to one-time errors.

The tray sizing program has been tested while communicating with the microcontroller with favorable results. The tray sizing program does need to be continuously run on Elizabeth's Mac.

Subsystem 6: AppleScript

An automatic image renaming script was created that renames all images in the output directory of the selected recipe by adding 98 to the last number in the image’s name. This needs to be modified to always set that number to 99, instead.

The autocrop script was tested with 109 test images. There were 8 failures, for an accuracy of 92.6%. Four of the failures occured when attempting to crop to a 3D object, a clarinet. Ignoring those failures, the accuracy of the system is 96.2%.

All AppleScript has been successfully installed on Elizabeth's computer.

Algorithm Failures

Image:Photo Gallery/failed test images/198120220009.0002.jpg Image:Photo Gallery/failed test images/E1900010005.0011.jpg

Low Contrast

Image:Photo Gallery/failed test images/201801060001.0001.jpg Image:Photo Gallery/failed test images/201600210015.0002.jpg


Image:Photo Gallery/failed test images/201700690001.0001.jpg

Test Results Summary

Testing Conveyor Belt without the eyelets: Moving Conveyor

Risk and Problem Tracking

Largest Risks at this time:

1. Time frame:

Mitigation Plan:The team is working outside of class time to ensure that all subsystems are completed before their due dates

2. Budget

Mitigation Plan: Currently Requesting more money from the RIT MSD office

3. Safety of Artifacts:

Mitigation Plan:'There will be a manual overrides for all moving parts in the system

Plans for next phase

As a team, we will:

(Use the individual 3-week plan template for this)

Home | Planning & Execution | Imagine RIT

Problem Definition | Systems Design | Preliminary Detailed Design | Detailed Design

Build & Test Prep | Subsystem Build & Test | Integrated System Build & Test | Customer Handoff & Final Project Documentation