Drones, Machetes, and Virtual Reality: 21st-Century Tools for Historic Preservation

by Dace A. Campbell, AIA, LEED AP
@DaceCampbell

Introduction
The climate is changing, and sea level is rising. Scientists estimate that ocean waters will rise by at least two meters this century, which could have a catastrophic impact around the globe. Among the nations and peoples most vulnerable to this coming change are those who make their home on the islands of the Pacific Ocean. Entire countries and cultures will disappear as beaches erode, villages submerge, and countless atolls and islands are wiped off the map. And with them, an entire history evidenced by architectural and cultural artifacts. But this tragic scenario doesn’t have to be the end of the story. In fact, with the right tools, you can help.

It seems that there is little that can be done to stop rising sea levels in the short term. However, as technology advances and becomes more accessible to professionals and amateurs alike, we have tools at our disposal to mitigate the impacts of climate change on our coastal artifacts. Specifically, aerial photogrammetry and virtual reality enable us to document artifacts at risk of submersion to support historic preservation efforts in the face of climate change.

I work as a Customer Success Manager at Autodesk, helping architects, contractors, and owners realize tangible value from their investment in Autodesk solutions. When I recently earned a sabbatical, I planned it with two goals in mind: get away with my family to a tropical island and engage in a meaningful philanthropic project.

I spent many months brainstorming and researching projects and places in which my family could participate in a meaningful project. Then, in summer 2017, when the Trump administration announced the U.S. would pull out of the Paris Climate Agreement, it clarified my focus. I decided to leverage my background in architecture and virtual reality technology to help mitigate the effects of climate change. Further research led me to find Kosrae, (pronounced: “kosh-RYE,”), Micronesia, where I executed a historic preservation project using photogrammetry and virtual reality to document culturally significant architectural artifacts being impacted by climate change.

Project objectives

The purpose of the project was three-fold:

1. Provide new data, points of view, and experiences of culturally significant artifacts to support effective decision-making about their preservation. This project supported Kosrae Historic Preservation Office in the discovery, clearing, documentation, and visualization of cultural artifacts to make decisions about how best to preserve them
2. Preserve cultural artifacts at risk of erosion/submersion due to climate change. The project empowered key stakeholders to visualize the past, present, and future state of prehistoric ruins with respect to sea-level rise.
3. Investigate feasibility, develop workflows, and apply technology including drones, photogrammetry, virtual reality, and augmented reality to historic preservation of cultural artifacts. The project is developing documentation to support a proposal by Kosrae Historic Preservation Office (KHPO) to UNESCO for consideration of World Heritage status of the Lelu Ruins.

Kosrae

I partnered with local leaders to execute the project in the winter of 2018, when my family lived for a month on the island of Kosrae, in the Federated States of Micronesia (FSM). Kosrae, well-known for its mountainous silhouette of a “Sleeping Lady,” is a stunning emerald island in the equatorial Pacific, southwest of Hawaii, northeast of Australia, and southeast of Japan. Kosrae boasts a reputation as the “Jewel of Micronesia” with lush mountain jungles, mangroves, and unspoiled beaches, and surrounded by thriving tropical corals. Almost all of Kosrae’s 6,000 inhabitants live along the island’s coast where they make their living from government employment, tourism, and subsistence farming and fishing.

Kosrae is also known for its prehistoric ruins, including the grand megalithic walls and compounds of the ancient political capital of Lelu and the remote mountainous temples in the prehistoric religious center of Menke. There are also 20^th-century artifacts along the beaches from World War II.

KHPO operates under the Kosrae Island Resource Management Authority (KIRMA), with a mission to preserve cultural artifacts as modern developments and improvements occur on the island. KHPO is engaged in several efforts to develop conservation areas in Kosrae, to protect ruins with the assistance of UNESCO World Heritage Organization, and to raise awareness of the value their unique artifacts offer the world. The staff at KHPO and KIRMA graciously welcomed me and my proposal to partner with them.

Challenges

KHPO, and the people of Kosrae in general, face major challenges with respect to sustainability.  According to KHPO Field Supervisor Swenson Thompson, “Most of our historic sites are near the shore, and vulnerable to climate change issues.”  Their cultural artifacts are threatened by regular “king tides” and climate change. King tides occur approximately seven days a month, and at high tide, historic sites are under water. At Lelu, the majority of the land is artificial fill, and most of it experiences regular flooding. These floods are exacerbated by rising sea levels. Artifacts that used to sit higher relative to the shoreline are now effectively at sea level. And it’s not just at the shore:  in the jungles inland, flooding from rivers is getting worse due to more intense storm activity, and sediment from river flooding is burying prehistoric artifacts more quickly than they can be uncovered and maintained. In short: the people of Kosrae are already losing their cultural artifacts.

KHPO also faces cultural challenges. Local land owners don’t always understand the value of the ruins on their family property. As they tend to focus on essential family needs over long-term benefits, some owners of the Lelu ruins are even seeking ways to profit from the sale of the rocks on their property. KHPO and the federal government struggle to convey their vision and the value of preservation to the land owners. Owners are unaware of the plans, and need good explanation before offering consent or approval. According to Ashley Meredith, Kosrae State Cultural Anthropologist, the information the general population takes in needs to be delivered more intuitively to bridge what is known and unknown. Further, without understanding, large groups of people can have difficulty making consensus decisions to take action. “If they [leaders] get a good picture of it, they should share that picture with others with less knowledge,” said Thompson.

Additionally, Kosrae can be an unforgiving environment in many ways. Paper media simply doesn’t last in the humidity, and digital media are not common. Access to internet and technology in general is poor compared to first world standards.

Finally, given Kosrae’s long history of exploitation as a colony, locals are rightfully skeptical of outsider expertise. As an outsider and subject matter expert, a phrase I was told was one that I experienced daily: “People don’t care how much you know, until they know how much you care.” While this is true in my “day job” at Autodesk, it was even more so working on this project.

Desired capabilities

KHPO staff are dedicated to their vision, and the historic preservation officials were quick to articulate desired capabilities for the people of Kosrae:

• Greater accessibility of experience for local Kosraeans: to extend experience of environments and artifacts to people who couldn’t or wouldn’t hike to remote conservation areas in the mountains
• Expansion of outreach off-island for researchers and historic preservation authorities
• To generate revenue for land owners with a well-communicated plan for funding preservation and operations of their ruins: “To simulate a management plan would be a very effective tool [for UNESCO] to understand the space,” says Ashley Meredith
• To connect Kosraean ex-patriots living abroad to their homeland resources
• To support tourism and education, in general
• To offer new vantage points for owners, giving them a clear view of artifacts without dense vegetation and geography obscuring their understanding of the sites as they plan their work to maintain stone ruins

Solution

Considering these challenges and desired capabilities, it became clear there was a place for my skill set, subject matter expertise, and modern technology to help. The solution I proposed included photogrammetry and virtual reality. For my stay in Kosrae, I brought a drone for aerial photography, multiple cameras for terrestrial photography, multiple mobile workstations and virtual reality hardware, and tablets for augmented reality. But the solution was not defined by hardware or the accompanying software – it was about experience. That is: aerial drone flights offer unique points of view of archaeological ruins not otherwise accessible from the ground, and virtual reality offers unique experiences of those same artifacts, to empower more effective decision-making in support of historic preservation efforts.

Aerial photogrammetry and virtual reality are about capturing and experiencing new points of view to make more effective decisions

Project sites

The ideal project site would meet (at least) three criteria: close to sea level; of cultural and historical significance on a global scale; and sufficiently cleared of vegetation to support aerial photography. Upon arrival to Kosrae, I evaluated several project sites. Some required clearing, but ultimately, I identified three sites on which to execute the work:

Japanese Radio Tower

The Japanese Radio Tower is a 20^th-century artifact from World War II, lying on the beach in ruins near the town of Malem. It gained fame for its use in the WWII movie, “Up Periscope” starring James Garner. Recent attempts to document the tower with LiDAR and other techniques failed, and KIRMA had prioritized this site for documentation. Located right on the beach at sea level, it is at significant risk of submersion from climate change, and tidal action has already knocked the tower off its foundation.

Menke Ruins

The Menke Ruins are found within Korsae’s “Valley of Faith,” in the Menke River valley near the center of the island, accessible by hike upriver from the town of Utwe. The ruins consist of prehistoric dwelling compounds and temples to Sinlaku, the “Breadfruit goddess,” and have only been rediscovered recently. Nearly 100 compounds are found along a 1-mile stretch, along with upper and lower temples. Because the ruins are about 300 feet above sea level, they are not threatened by sea-level rise. However, climate change has increased precipitation in the mountain jungles, and flooding has washed so much sediment around and into the compounds that the walls are of unknown height below the current jungle floor.

Lelu Ruins

Lelu was the prehistoric political and administrative center of Kosrae, and built up over centuries between 1250 and 1850. Its Megalithic stone walls reach over 20 feet in height, built of columnar basalt and coral construction of a complexity and architectural significance that rivals Nan Modal on the neighboring Micronesian island of Pohnpei. Nan Modal is protected by UNESCO’s World Heritage Organization, and KIRMA and KHPO are proposing that Lelu share that protected status. Historically, Kosraean peasants would access Lelu by canoe in narrow canals at high tide to bring tributes to the high chiefs who dwelt there. Today, flood waters regularly spill over the canoe landings during “king tides,” limiting accessibility to and through Lelu. There’s no doubt the ruins will be severely impacted by rising sea levels, and the ancient city is at risk of being submerged or damaged by erosion from increased flooding.

The Process
In the weeks leading up to my travels to Kosrae, I procured photography, computing, and virtual reality hardware tools from multiple vendors and sources, including: a DJI Phantom 4Pro aerial drone, Canon EOS Rebel DSLR camera, Apple iPad, Samsung Galaxy mobile phone Lenovo ThinkPad, HTC Vive, HP Omen, and Samsung Odyssey. I also installed Autodesk Recap Photo, Autodesk 3ds Max, Autodesk Max Interactive, AutodeskRevit Live, Steam VR, 3DR Site Scan, Visualive3D Mobilive, Geotag Photos, and Trimble SketchUp software applications.

With these tools, I developed and tested workflows to collect data, document it with photogrammetry, and experience it with virtual and augmented reality before traveling to Kosrae. I used my own house and neighborhood plus referenced drawings from the book “Prehistoric Architecture in Micronesia,” by William N. Morgan as source data to create a digital twin of my front yard. I then flooded that virtual environment with rising sea levels, populated it with ancient Kosraean dwellings, and overlaid those virtual dwellings onto the physical front yard. I was satisfied with the with the workflow test results by the time I had to pack up and get on a plane.

   ,   f.

Although the overall workflow I developed was rather complex, it breaks down into the following steps:

1.      Collecting images with a drone

To photograph the sites from the air, I used a DJI Phantom 4 Pro aerial drone piloted remotely with DJI Go 4 software on an Apple iPad. This is relatively straightforward, but for anything except the simplest of drone flights, it’s recommended to use legitimate “mission planning” software. I chose 3DR SiteScan, which has a simple interface that enables you to plan and fly sophisticated flight patterns. To plan flights, it’s helpful to first have a good top-down (“ortho”) photo of the area you want to document. None of the sites in Kosrae showed in the low-res satellite photos I had available, which merely revealed nothing but tree canopies and clouds. To generate an accurate ortho photo, I flew a high overhead pass (between 300 and 400 feet) over the entire area and processed those photos to establish a current, high-resolution background image.

Prior to arrival in Kosrae, my assumption was that the Lelu Ruins were relatively clear of vegetation, as documented in Morgan’s book on Micronesian architecture. However, photos from that book are 30 years old, and my first overhead flight confirmed what I feared from our initial treks through the ruins: thick vegetation covered nearly every square meter of terrain, wall, and canal. Brush, ferns, coconut trees, breadfruit trees, Nepa palms, and great Banyan trees formed a dense jungle canopy over the ruins. This vegetation would need to be cleared to have any realistic opportunity to photograph the ruins from the air, or from the ground.

Expectation vs. Reality:  Representations of the ruins in literature on the left, dense jungle growth on the right

I processed the photos from the initial drone flight, created a point cloud to take measurements of the tree canopy, and generated a mesh from the dataset to walk through immersively in virtual reality (VR). Besides confirming that there was extensive coverage from vegetation, the VR experience offered a key insight: using the exocentric, “mini-map” mode offered in Autodesk Revit Live, I observed that the trees in a particular area were less dense, the brush was thinner, and a few spots of the canal were visible from the air. This area was in the vicinity of some of the more spectacular stone walls, and I concluded that it could be cleared with less effort than other areas of the ruins. This insight was not at all clear from visual analysis of the ortho photos, even at high resolution; VR enabled me to perform a spatial analysis to reach this conclusion.

I pitched a proposal to KHPO that we clear the Lelu Ruins in a specific region up to two acres (10,000 square meters) in area. The intent of the proposed clearing was to reveal specific stone walls with varying structural composition or archaeological significance, and I developed flight plans to best capture their likeness in and around the significant trees that would remain. KHPO agreed to my proposal, and it was time to assemble a labor crew.

Original satellite image, photo locations from initial aerial flight, new ortho photo, and snapshot of VR experience

To clear the site, we enlisted the assistance of Kosraean prison labor, compensated with extra food for them and their families. My family and I worked side-by-side with them and staff from KIRMA and KHPO, swinging machetes and hauling debris in the oppressive heat and humidity for several days. With one week left to my stay in Kosrae, I had perfected my machete swing, and a select portion of the Lelu Ruins were cleared enough to photograph from the air. At Lelu, I collected nearly 1000 photos of the ruins across 9 automated aerial flights, plus another nearly 900 photos from manual aerial flights.

With the help of prison labor, we cleared enough vegetation to expose various compounds to aerial photography

2.      Collecting images with a camera and geotagging

To supplement the automated and manual aerial drone flights, we also photographed details of the stone walls with terrestrial photography using a Canon EOS Rebel DSLR. This is a relatively straightforward process, but does require careful attention to lighting and aperture settings, as blurry photos will not process in photogrammetry.

My DSLR does not support geotagging, but some photogrammetry apps require such meta-data be included in each image for processing. To overcome this, I used the Geotag Photos app on my Samsung Galaxy mobile phone, which tracks and associates your location (via GPS) to the local time to create a GPX file, which can then be used to generate geolocation information for each photograph. If the clocks on the phone and camera are synchronized, this is relatively simple to do. We took over 800 terrestrial photos of the Lelu Ruins, plus hundreds more of the other sites, and it was easy to geotag them in batch processes.

3.      Uploading images for photogrammetry

Once the photos are taken, they had to be uploaded to “the cloud” for photogrammetry processing. To process my pictures, I relied on both Autodesk Recap Photo and 3DR Site Scan. Uploading and processing is a relatively basic step, assuming you have sufficient bandwidth. Kosrae does not. With over 6000 photos taken across all the projects, uploading files for processing proved to be one of the largest challenges of the project. We were provided with the broadest bandwidth on the island, but photos that should take just 1-2 seconds to upload under modern bandwidth standards would take 3-4 minutes each, if the upload didn’t time out. To make matters worse, the internet for the entire island frequently goes out in the wind or the rain, and we executed this project during the rainy season. At times, the internet would be out for four consecutive days, severely limiting our ability to upload and process photos.

Our experience in Kosrae is not an outlier or an exception in this regard, and the computer industry should take note. No doubt there are tremendous benefits to cloud computing, streaming, SAAS business models, etc. However, without sufficient bandwidth, this trend of moving away from local computing could serve to further alienate Third World countries without sufficient bandwidth to upload and download data from the cloud. We cannot afford to leave the most underserved populations on earth even further behind in this regard.

It’s also worth noting that just because one can take thousands of photos of a site, doesn’t mean that one should. Nor can they all be used effectively in a single photogrammetry process. There are limitations and implications for the number of photographs submitted for processing, as follows:

• 20: Minimum limit of images to process
• 300: Maximum limit of images to process in Recap Photo “object” processing
• 500: Limit of 3DR Site Scan use of background Recap Photo, switches to Pix4D if higher
• 1000: Maximum limit of images to process in Recap Photo “aerial” processing
• 2500: Maximum limit of images to process in 3DR Site Scan

Awareness of these limitations is vital, and they demand a carefully planned work-breakdown structure. Managing thousands of photos and a thoughtful structure requires good “digital hygiene” of file management, naming conventions, discipline, and record-keeping. The significance of this effort cannot be overstated.

Artifacts from the photogrammetry process include point clouds and surface geometry:

• 3DR Site Scan does an excellent job of creating point clouds from automated aerial flights planned with its tools, and further offers a suite of tools for analysis and measurement of the point cloud data. It also supports export of OBJ and RCM files, which can be consumed by Autodesk tools for editing.
• Autodesk Recap Photo creates a textured surface mesh via “aerial” and “object” processing methods, and each has strengths and weaknesses depending on the nature, quality, and quantity of photographs uploaded. It also offers a full suite of tools to make basic edits like measuring, deleting surfaces, filling holes, and decimating the mesh for export and use in other tools.

Point clouds as an artifact of photogrammetry, depicting the Lelu Ruins from the ground and the air

4.      Building a digital model to scale

Photogrammetry was the primary “reality capture” data source for this project, but for a specific aspect of the virtual reality and augmented reality experiences, I wanted to supplement this with full-scale models of prehistoric Kosraean dwellings. To do this, I referenced drawings from Morgan’s Micronesian architecture book and used Trimble SketchUp – an easy-to-use modeling program for architectural forms and components. While SketchUp doesn’t natively embody meta-data associated with Building Information Modeling (BIM), this was not necessary for the basic geometric representation I needed for the project.

Source imagery and model in-progress of prehistoric Kosraean dwelling common to Lelu

5.      Processing models

With digital models of the environment from Autodesk Recap Photo and 3DR Site Scan, supplemented by models of dwellings from Trimble SketchUp, I imported OBJ and SKP data into Autodesk 3ds Max, an application commonly used for modeling, rendering, and animation. In Max, I aligned geometry, made minor edits, and exported FBX files for consumption in Autodesk 3ds Max Interactive.

In Autodesk 3ds Max Interactive, I aggregated model geometry, defined navigation meshes, refined materials, and with the help of colleagues at Autodesk I generated interactions to be able to toggle on/off the dwelling units as well as raise and lower the water level. I then compiled and deployed stand-alone executable files for consumption in virtual reality using Autodesk Revit Live.

Autodesk’s mission is to help people imagine, design, and make a better world, and with this project I could do just that. All of these tools by Autodesk – Recap Photo, 3ds Max, 3ds Max Interactive, and Revit Live – are creation and visualization tools that enable professionals and amateurs alike to capture, create, compute, and consume data. I’m confident in saying that anyone interested in applying photogrammetry and virtual reality to historic preservation can use these tools as an integral part of their workflow.

6.      Virtual reality

A prime deliverable from this project is an immersive virtual reality (VR) experience of the cultural artifacts for professional staff and local land owners alike. VR offers many things, but in the context of historic preservation it supports effective decision-making through the experience of a space. VR is an immersive, interactive, and intuitive medium, placing a participant into an environment to help them understand, empathize, and form a consensus opinion with other participants about what action to take, based on shared experience. Traditional visualization offers representations that are open to interpretation, which means they are also susceptible to misinterpretation; in historic preservation such misinterpretation can increase the risk of confusion, poor communication, delays, rework, or worse: damage to the artifacts themselves.

Scaled-down traditional representation vs. full-scale immersive virtual reality experience

For this project, I developed VR experiences to be consumed in a few ways. Primarily, executables deployed from Autodesk 3ds Max Interactive were created to run Autodesk Revit Live using HTC Vive head-mounted display and controllers. This common VR set up allows participants to wander freely within a room-scale space and use virtual navigation tools to explore a large-scale environment. Participants can walk around egocentrically (first-person point of view), turn prehistoric Kosraean dwelling on and off to sense how the ruins appeared in the past, and raise the sea level to sense the impact of climate change in the decades to come. Participants can also explore the ruins exocentrically, manipulating the environment as if it were a scaled-down model placed on a table.

The HTC Vive system I used has video cables and external tracking base stations that limit a participant’s mobility to a “room-scale” space a few meters across. I also tried a Samsung Odyssey “Windows MR” head-mounted display cabled to a portable HP Omen backpack computer. Free of tethers, and with the “inside-out” tracking capability of the Odyssey, a VR participant could theoretically wander at will throughout an expansive physical space.

The VR executable files also can be run on a desktop workstation without immersive VR hardware to create a “desktop VR” experience. Like a video game, participants can also move around the environments in real-time with a mouse and keyboard, while also toggling dwellings and adjusting sea-level height. This “desktop VR” variation makes the virtual ruins far more accessible to a broader audience without the immersive VR hardware.

7.      Augmented reality

In addition to consuming the data immersively in VR, I created an augmented reality (AR) experience using the Visualive3D Mobilive app on an Apple iPad. Unlike VR, which completely immerses a participant in an artificial environment, AR serves up spatially contextual data when and where it is needed. By overlaying a digital model onto the physical environment, the participant augments their view and experience of the real world.

For this project, I overlaid the model of the prehistoric Kosraean dwelling onto the real world, enabling land owners to visualize and experience the ancient wood and thatch structures at full scale, in context of the stone ruins on their property. They shared that seeing the virtual past overlaid onto their property was a captivating and eye-opening experience.

Results

Having spent considerable time in Kosrae clearing the sites and documenting them, I successfully performed photogrammetry and created several VR and AR experiences, toured by dozens of participants including government officials, field staff, archaeologists, land owners, and students and staff of the College of Micronesia. While the environments they experienced were works-in-progress, the impact and the feedback I received from their experience were quite positive.

Virtual Japanese Radio Tower (exterior and interior)

Virtual Menke Ruins (aerial view and altar detail)

Virtual Lelu Ruins: with prehistoric dwellings, wall detail of Kinyeir Fulat, wall detail of Finbota, and future flooding

Despite all the Third World limitations with bandwidth, internet outages, labor, and overall frictions of cross-cultural communication, the effort was a smashing success. Kosraean participants in the project had many positive things to say about the experience. I interviewed several of them, including administrators, an anthropologist, field staff, and land owner of some of the ruins.

Blair Charley, Administrator of KIRMA, had the following to say:

“I was very impressed with the technology as it let me look around as if I was actually there, on site with the ability to include the original, ancient Kosraean houses, to come a bit closer to seeing and imagining how the Ruins were settled back then. The capability to project sea level rise and climate change into the virtual space, also lets me experience a near life like scenario in 50 years’ time. I find it just very useful and wonderful to use and experience for the first time. 

“Drones, photogrammetry, and VR are very useful tools, not only to KHPO work but other work involving environmental conservation, infrastructure development and planning, climate change adaptation and relocation, etc. Our biggest concern in terms of survival and for future generations is Climate Change, and simulating the effects and rise of sea level through VR can really make an impact. Viewers/users will experience the near life like scenarios of rising sea level and the impact this may have on existing structures. Showing how sea level rise may inundate development should support current decision making to address impacts of climate change. This can be the same with conservation areas.

“Currently we anticipate getting the Lelu Ruins recognized as a world UNESCO site. But ambitions to have this goal realized really depends on the landowners. We support this site getting further recognition, and we can make the case by using VR to enhance a reader’s understanding of the Ruins and how modern technology in support of or could further support its case for preservation.”

Andrew Standon, Director of KHPO, immediately recognized the value of having drones in their stable of tools, and requested a training demonstration for his staff. He expressed strong interest in leveraging drones to support their ongoing field work.

Ashley Meredith, the Kosrae State Cultural Anthropologist, expressed a strong desire to learn more about the technology. ““Before, I was challenged by not tapping into my spatial faculties; VR may change that.”  VR can lower the barrier to experience, enabling Kosraeans to “be able to see things they won’t otherwise be able to.”  She saw VR as an excellent tool for getting data into people’s hands, as an effective way to communicate with off-island researchers and peer professionals, and to connect Kosraeans living abroad to their homeland resources.

Swenson Thompson, Archaeology Field Supervisor of KHPO, offered the following:

“This is the first time I’ve experienced VR technology and how it can greatly improve our work here. The ability to ideally experience exploring through a site that you’ve never been to is great. Drones, photogrammetry, VR, and AR will be very helpful to our office (KHPO/KIRMA) to build 3-dimensional photos or maps to be used and may also be useful to other units under KIRMA such as Forestry, GIS, or even the Marine unit ‘monitoring.’ Kosrae could use these tools to help preserve the Lelu Ruins by producing advertising materials or video/photogrammetry images as awareness products and be displayed globally for tourism and economic development purposes. If only I owned one, I could do my work easily and enjoyably. It has been fun and informative in terms of what contemporary technologies are available to assist conservation, preservation, as well as archiving the resources.”

Finally, Salik Waguk, owner of the Menke Ruins, observed that VR offers déjà vu in another world. For him, it was good to focus on just the stone ruins after vegetation was virtually and physically cleared: “VR gives me a clear view of the place without the vegetation or hills.” He wants to use VR to plan his work to maintain the ruins, allowing him to see the place as a clear area. To him, drones offer him a unique experience allowing him to view his work from above. He expressed strong interest to try it again.

Conclusions and Next Steps

Based on my experiences working on this project, I’ve identified a handful of best practices and lessons learned when working with photogrammetry and virtual reality.

These include “pro tips” for planning and executing automated aerial photogrammetry such as with 3DR Site Scan:

1. Use a spotter, especially where you can’t maintain line of site with the drone
2. Start with high-flight at 300’-400’ to establish a base ortho photo
3. Practice and perfect drone flight patterns to re-fly final flights automatically
4. Fly all flights at same time of day for consistent lighting and shadows
5. Fly in low wind and in overcast skies or hazy sun to minimize dark shadows and glare on water
6. Carefully plan flight times to optimize battery use
7. Record times and numbers of photos in areas you know you can delete
8. Measure a known distance on the ground for reference
9. Bring an umbrella to keep your tablet dry (rain) and cool (sun)
10. Bring along spare cables, adapters, batteries, and propellers

BONUS: Wind gusts are unpredictable, and trees eat drones for breakfast… be flexible and stay alert!

Additionally, I discovered multiple use cases for VR in a photogrammetry workflow. One preconception I had was that VR would be used to visualize the final model, after photogrammetry was complete. This was certainly true, and validated my entire effort. However, I also found VR to be quite useful during the photogrammetry process, as follows:

• To review low-res models created from an initial high-flight, to be able to identify areas of interest and plan additional flights to document them
• To quality-check in-progress models, to be able to identify gaps in the data captured and plan supplementary drone flights and photos to fill those gaps
• To compare models of an environment processed from photos taken from different flights, whether from different flight patterns, or various times, days, or weather

Overall, an immersive VR experience proved invaluable throughout my time capturing photos and planning flights. In this way, photogrammetry and VR became co-dependent and reiterative processes, each one informing and improving the other at the next iteration.

Work on this project continues today, and there are three major next steps ahead. First, I continue to work on the VR experiences, to move past “proof of concept” or interim models to create more final, complete experiences. I continue to sort, group, and prioritize processing of more than 6000 photos, revise work-breakdown structures, and process and aggregate newer, more detailed models. These models contain 150 million polygons, which need to be simplified to models of less than 8 million polygons for real-time viewing.

I also intend to share the data set back with the people and government officials of Kosrae. In addition to all the photos and derivative models, I intend to provide access the “desktop VR” experience of the ruins from a web site to make it broadly accessible to Kosraeans and off-islanders alike. This final experience may also include video, imagery, and text-based descriptions supplementing the VR experience.

Finally, I’ve offered to support KIRMA and KHPO’s grant applications and project proposals with this technology. Whether that is a proposal to UNESCO to grant World Heritage status to the Lelu Ruins, or grant applications to procure their own photogrammetry and virtual reality technology, I intend to continue working with Kosrae to put these tools into their hands to preserve their cultural heritage.

The climate is changing, and sea level is rising.

You can help document and save cultural artifacts before it’s too late.

[2018.05.24 update: The VR experiences can be downloaded as executable files here.] 

[2019.05.01 update: A 4-minute video describing my experience and the impact it had on my career can be found here.]

Acknowledgements
This project was made possible by an army of supporters. Government organizations, companies, and individuals from both the USA and FSM donated their time, attention, authority, and resources to make it happen. I would like to sincerely thank the following for their support:

Kosrae Island Resource Management Authority (KIRMA)
Blair Charley

Kosrae Historic Preservation Office (KHPO)
Andrew “Stan” Standon
Swenson “Swen” Thompson
Ashley Meredith
Andy William
Austin Albert

Kosrae Forestry Department
Maxon Nithan
Lance Nithan

Menke
Salik Waguk

FSM National Government Historic Preservation Office
Augustine Kohler

Kosrae State Government
Carson Sigrah, Lt. Governor

National Park Service
Office of Insular Affairs
Lelu Ruins Association
Kosrae Department of Public Safety
Christophe Sand, Archaeology Institute of New Caledonia

WiFi Access
College of Micronesia – Federated States of Micronesia
Green Banana Paper Company
Kosrae Pacific Resources for Education and Learning (PREL)
Kosrae Nautilus Resort

KnowledgeWell
Aaron Smith
Felicia Beardsley

Autodesk
Sebastian Casallas
George Hatch
Kellan Hays
Hans Kellner
Jarrod Krug
David Menard
Mike Mizuno
Chip Weatherman
Tim Young

Hardware and Software
3DRobotics
Lenovo
nVIDIA
Visualive3D
Hewlett-Packard

Supporters
Joel Pennington
Rama Dunayevich
Dave Schlesinger
Joe Travis
Sandy Gilson
Dennis & Terri Campbell
Dean & Karen Campbell
Chris & Tina Arjona
James & Karen Tanney
Todd & Michelle Whitehead

Finally, a word of sincere appreciation to my family for accompanying me and participating actively in this project, even as I plucked them from their cozy home and dropped them into the steamy jungle for a month. While I’d like to think the hiking, paddling, diving, beaches, and coconuts around Kosrae made it worth their while, I know they sacrificed a lot to work, clear vegetation, take photographs, and sit through “boring meetings with grown-ups” nearly every day. Thank you, Susan, Arlan, Corban, and Declan!