1/21/2020 White Paper Write Up

Austin Sullins

AT 419

            Intoduction:

All of last semester my group work nonstop trying to find a more efficient way to 3D map all sorts of buildings. We flew over 10 flights gathering data and have built over 5 3D models. With most of our data and research done we are starting to prepare ourselves for the publication aspect of this project. This intern is by far the least interesting part of this entire project, BUT the absolute most important! This is the part of the project that all of you who are reading this will see. This is the part that you will judge us upon. This is the part that will be the deciding factor of whether we get a job in this industry or not, and that is why we are putting in so much time and energy into the decision of doing a white paper or doing a pier research paper.

            Some of the main differences between a white paper and a peer review is what you are trying to accomplish! If you are trying to do more of a sales and pitch then a white paper is what you would want to go with. If you are looking to have someone read your material to lean or teach with then a peer review would be a better fit. I personally have written a few white papers so I believe that is the path that my team should go down. However, after the research I have done and what we are really wanting from this project the peer review is starting to sound like it could be a good path to go down and here is why.

           

https://catsr.vse.gmu.edu/pubs/S2DFS_Con_Ops.pdf

https://www.sensefly.com/app/uploads/2018/05/Comparing-workflow-and-point-cloud-outputs-of-the-Trimble-SX-10-TLS-and-senseFly-eBee-Plus.pdf

https://www.frequentis.com/sites/default/files/support/2018-02/37_ATM_Whitepaper_drones_1017.pdf

           

            First thing I notice right off the bat is that the structure is completely different. It reminds me of military writing how everything is flush to the left margin unless it is an extended note of that previous statement. Another thing is the utilization of diagrams and charts. This is something used mostly for data in peer review and not really a thing in normal literature. The reading style is very easy to follow and flows with the process that they are trying to complete. I do believe that this is a good process for certain things that you are trying to get done. This style would work a lot better than a normal paper, but I don’t believe this is the path our group should follow.

                       

Conclusion:

Yes, we are trying to “sell” ourselves to companies and possible employers, but I believe that isn’t what this project is about anymore. I believe that is has grown more into testing and stretching the bounds of what this industry knows and is comfortable with! The thought of accurate 3D mapping taking a matter of minutes and as few as 5 pictures seemed impossible to us and, from what I’ve gathered from talking to companies, the industry as well. Yet, our team has been able to cut down this process by over half and we aren’t even experts in everything we need to be to make the process the fastest it could be. Along with that we are asking questions that haven’t really been asked before. I had a conversation with multiple people from ESRI about our project and what we need to make this process more efficient and the people I had talked to had never thought about the things our team was doing. They had never thought about taking an actual CAD blueprint of a building and overlaying pictures on top of it. They never thought to look into how this could turn building inspections into a process that takes minutes instead of days. All of this added together is telling me that a peer review paper is the way this group needs to go to open more doors up for not only ourselves, but for this research as well.

Work Cited

“Cultural Confusion: White Papers vs. Peer Review.” Digital World Biology, 30 Oct. 2017, digitalworldbiology.com/archive/cultural-confusion-white-papers-vs-peer-review.

Dorard, Louis. “The Problem with White Papers.” Medium, PAPIs Stories, 17 Sept. 2018, stories.papis.io/the-problem-with-white-papers-d82c3678f9e7.

Kelly, Jacalyn, et al. “Peer Review in Scientific Publications: Benefits, Critiques, & A Survival Guide.” EJIFCC, The Communications and Publications Division (CPD) of the IFCC, 24 Oct. 2014, www.ncbi.nlm.nih.gov/pmc/articles/PMC4975196/.

Ordway, Denise-Marie. “White Papers, Working Papers, Research Articles: What's the Difference?” Journalist's Resource, 4 May 2018, journalistsresource.org/tip-sheets/research/working-papers-research-articles/.

Purdue Writing Lab. “Purpose and Audience // Purdue Writing Lab.” Purdue Writing Lab, owl.purdue.edu/owl/subject_specific_writing/professional_technical_writing/white_papers/index.html.

https://catsr.vse.gmu.edu/pubs/S2DFS_Con_Ops.pdf

           

https://www.sensefly.com/app/uploads/2018/05/Comparing-workflow-and-point-cloud-outputs-of-the-Trimble-SX-10-TLS-and-senseFly-eBee-Plus.pdf

https://www.frequentis.com/sites/default/files/support/2018-02/37_ATM_Whitepaper_drones_1017.pdf

GIS Day

We started GIS day by putting up the display of our poster.  I was surprised that the UAS posters represented over half of the posters being presented that day.  The other posters were very interesting and had a large range of topics.  One that I found intriguing was on snow removal and how they sectioned them out.  Below you can see what our group's poster looked like on display.

The arrangement that the posters were set up in was very odd.  They seemed to be out of the way and presented as a second afterthought.  An alternative setup that could be better is placing the posters where you walk in and the seating area after the posters.  Setting up like this would help draw the guests to the information on the posters first before taking a seat.

Several speakers from different parts of the industry came and spoke about how GIS is used on a day-to-day basis within their respected fields. The keynote speaker, Carly Sakumura from Maxar Technologies, explained how their personally owned satellites are used for various fields and industries. She also explained the process of how the satellites collect pictures and clean them up to make them easily viewable.  Below is a photo taken during Carly Sakumura's keynote talk.

Other speakers, such as Zachary Brecheisen spoke about the quantifying arid-region agricultural development in Peru using Landsat. Zachary has been doing this research for a few years and has figured out some of the driest areas in the region focused on in Peru.

Another highlight speaker was Christina Hupy, who explained the value of open source GIS in higher education. Christina explained that open source GIS is useful to teach to students because students are then able to utilize free GIS software without having to buy licenses from Esri. This allows students to develop their skills on their own time and gives them another edge when competing for jobs in the industry. The image below is Christina Hupy discussing open source GIS.

GISDay Poster

Week of October 29th

With GIS Day coming up at Purdue University, a poster highlighting 3D modeling in Multipatch and Pix4D is being created to submit for the event. Last week, each team in the class showed their rough draft poster to be critiqued and improve. Below is each students suggestions and how the poster was revised based off said suggestions.

Key: 

• Suggested change to the poster
• Poster change made

Hupy:

• Rename Pix4D
• Changed Pix4D name to photogrammetry software
• Change arrow color
• Changed from a yellow arrow to a black box outlining the picture of background noise
• Explain photogrammetry initially and what kind of software
• Explained how photogrammetry software is used to generate structure from motion, multi-view-stereo 3D models

Jasper: 

• abstract center
• Abstract was placed in the top left 
• Move pros boxes down for more space 
• All caps on abstract. SATT logo messed up
• Fixed the abstract so it is no longer all capitalized and SATT logo is no longer distorted
• Adding space down the center of the page
• Centered the page up better for uniformity 
• Bottom line up is offset
• Bottom line is no longer offset

Sam:

• Black border around the poster
• Did not end up doing a black border around the poster
• Label model figures
• Did not end up labeling model figures as it is explained within the text next to each model

Alan:

• Box for round barn zoom up
• Added a black box around the round barn zoom up
• Make it flow better
• Lined up Multipatch and Pix4D so the poster flows better and is easier to understand

Luke Hull:

• Switch the whole poster
• Made the poster readable form left to right for better flow
• Better layout
• Improved the layout for better flow

Ryan:

• Better layout. Each one is a couple pixels off
• Improved the layout for better flow. Improved the designs on the poster so they line up evenly
• The gold bar at the bottom of the page
• Fixed the gold bar to line up evenly
• Remove text boxes or line them up better
• Ended up removing most text boxes for a cleaner look

Connor:

• Balance the poster
• Balanced the poster for a cleaner look and better flow
• Just have pros and cons of both
• Ended up splitting the pros and cons by which software is being talked about for easier understanding
• Center multipatch in the poster
• Ended up putting Multipatch at the top of the poster rather than centering it
• 
  Highlight the area zoomed in on for building
• 
  Put a black box around the zoomed up picture to make it easier to notice
• Question and overall title contradict each other
• Kept the title and question the same
• Abstract is dense
• Kept the abstract the same, not able to condense since it is covering a lot of information 

Tim:

• No text boxes
• Removed text boxes for a cleaner look
• Format top down
• Ended up formatting top down

Luke Mklar:

• Boxes don't need the borders
• Removed the text boxes for a cleaner look
• Photos in the center doesn't work
• Left the photos in the center and evened them out on each side

Jarrett:

• Messy and inconsistent
• Cleaned up the layout for a cleaner look and easier to understand
• Line up the text boxes better
• Removed text boxes for a cleaner look
• Use the Purdue official colors
• Purdue official colors were too dark, used a slightly lighter color
• Flip the top 2 so multipatch is on the right
• Multipatch is on the top while Pix4D is directly below.
• Arrow is cheap. Use inset lines instead
• Added black box highlighting the zoomed in photo instead of the arrow

Below is the final poster that will be submitted for GIS Day. Most suggestions were taken into consideration vastly improving the quality of the poster.

GIS Day Poster Rough Draft

Introduction:

We have prepared a thorough rough draft for the poster we are presenting on GIS Day.  The abstract was positioned in the top left corner so the viewer focuses there first.  Below the abstract, we detailed the pros and cons for each method of creating 3D models.  Then next to this we have created a table to illustrate the different times that each phase took to complete.  These phases include post-processing, processing, and data collection.  After this we included the 3D model results from the Multipatch and Pix4D.  Next to the models we made sure include brief descriptions of the software and how it was used. The features of the poster at the bottom are still in the process of being created and we have used placeholders for what it could look like.  This is one of the final deliverables we need to finish before making the final poster.

Advancements:

To ensure the best possible accuracy and quality from Pix4D, the model undergone post processing. Ground control points (GCP's) were added for accuracy and masking was done to reduce the clutter. masking was done to the 3D model to get rid of background noise.

Pix4D Model Before Post Processing 

Reducing the background noise helped improve the overall quality of the model, however, it took an extra 28 minutes which is a crucial amount of time.

Pix4D Model After Post Processing

The office building in multipatch was also improved to better the overall quality. One continuously arising issues with multipatch is that it is glitchy. Shown in the picture below, the red "editing" color on the back side of the roof would not go away. Furthermore, it sometimes does not allow for the multipatch to be edited.

ArcGIS Pro Multi-Patch Model

Work Done this week:
Our team increased our level of understanding in CAD software, and the post-processing of 3D models and we recently found out that Multipatch may have post-processing that allows for features that cannot be added in ArcPro. There will be more to come on that as we figure out the CAD software, stay tuned.

Weekly Overview:

We made large strides this week in our overall knowledge off 3D modeling and the multiple different software available to make the process more efficient. My personal knowledge of Fusion 360 is increasing exponentially. I am following a course on the program from the fusion 360 website and also getting tutored on how to use this program along with CATIA.

Week of Oct 15 2019

Introduction:

This week involved more advanced features in Multi-patch and the beginning of the creation of everything we will use for GIS Day at Purdue. We collected more data at the Purdue Beef  Unit, this data is non-rectangular so it will challenge the Multi-patch software and our team to see if creation of these models are still easier than the standard 3D modeling. For GIS day, where numerous industry professionals will come to Purdue to look at the work that researchers are putting in, we started creating a poster portraying what Multi-patch can do.


Adept Skills in Multi-patch:

At the Purdue Beef Unit, we focused on non-rectangular and buildings with convex roofs. This will allow us to work increasing our skills in Multi-patch creation and to analyze if these intermediate buildings are still done easier versus doing them in a basic 3D model in PIX4D

Convex Roofs in Multi-patch

The roof was very easy to separate and pull up the convex sections, but our one failure of the data collection was that we did not gather the dimensions yet, so the convex roof points were done on a freehand basis.

Convex Roofs in Pix4D

The pro of the creation of the Pix4D model is that the roof overhang on the sides of the building is shown in 3D versus just being an imported photo. The downside is that the noise on the model has messed up the sides and the roof, leaving it looking patchy at points

The timing of the collection and the processing still points to Multi-patch being superior. Below is a breakdown of the timing.

GIS Day Poster Template:

The main things that we tried to focus on for this template creation were based off of how the human eye will normally interpret things. Western language speakers will naturally read left to right, then top to bottom. Using that information, we put our research question and abstract first, so immediately the main point would be read. Right under that we put our comparison of each of the software packages, this puts all vital information nearby. On the right of that section, Multi-patch and PIX4D analysis are listed. Again, since the human eye reads top to bottom, the Pros of Multi-patch will be seen first and interpreted as superior. Below that section are the section for importing and exporting from Multi-patch. This is listed as importing on the left and exporting on the right to flow easily with the reader. More work will take place as time goes on, but this is the start of presenting our data.

Class Poster Creation:

Our class was assigned  to create a poster to be placed in the hallway of the Purdue Niswonger Building. We were allowed to create the poster using any topic and focusing on any audience of our choosing. Once we decided that our audience would be parents, prospective students, and current students, we decided that the topic should show off why Unmanned Aerial Systems is an upcoming and relevant field. Because of this, we were able to use data collected from our capstone project with 3D modeling along with a focus on High Resolution Imagery, GIS Data, and Search and Rescue.

My Contribution:

For the design of our new class poster I was vetted and selected as project manager, in charge of the entire Unmanned Areal Systems capstone department. I was less of the small hands on stuff and was tasked with designing, finalizing, and printing. Keeping everyone on task was quite easy, because I just let them know that the sooner we get this done the sooner we could leave for lunch! We broke off into two groups, one to go take the fleet picture, and one to build the actual poster. The overall construction of the poster went smoothly once we decided on the poster outline it was smooth sailing.

White Paper Intro

3D modeling using UAS platforms is a relatively new field and little research has been done in best practice. Although research done by Stătescu and others have looked into software packages best suited for the job and which UAS platforms work the best, none have come to concrete conclusions. Several studies suggest that Pix4D is the best software to use when 3D modeling structures, however, this can be challenged. 3D modeling using ArcPro may be the most efficient and accurate way of 3D modeling. The question to look into is if ArcPro is an overall better software to use than Pix4D. This will be determined by how long it takes to capture the pictures in the field, how long it takes to process, and the overall end result. 

When comparing software it is important to note that each will have their own problems that need to be considered.  It was found in a study by Arefi and Alidoost, that all software will have errors but it was found that most software experience similar problems.  Software like Pix4D creates larger gaps between surfaces due to the angles that the images were captured in. The processing time with Pix4D was found to be the most efficient of those compared.  Other software that were included in the study were 3DSurvey, Agisoft Photoscan, and SURE. Each software processed 300 photographs to create seperate 3D models. The longest amount of time to process the data was roughly 10 hours by 3DSurvey.  Pix4D was able to process the data with negligible errors in only half the time at 5 hours. 

With the limited amount of research that has been done there are very few ways to get an actuator 3D model of a building to overlay the images in multipatch. These processes take a long time even for someone who has been doing it for months. There is a need and a demand for simple, fast, and effective way to be able to create a CAD model within minutes to be able to work with for a 3D model. One single software has yet to be invented to do all of this so the combination of softwares such as, Scan-To-CAD, AutoCAD, ArcGIS Pro, ect. will need to be used in a synchronized system to accomplish the desired results.

While ArcPro allows for more advanced input, the software also allows for more advanced outputs. Once a multipatch has been created, you can export a multipatch as a COLLADA file. Once exported as a COLLADA file, it can be saved to allow easy building of a scene versus just one building. COLLADA also saves the GPS location of the object created, allowing for drag and drop easiness in creating a scene. COLLADA files also allow for greater detail to be saved in the file, so looking at the file through Microsoft’s Hollowlens or 3D printing would keep greater detail than the average OBJ file.

Story map

The link above is an Esri Story Map showing the sequence of comparing Pix4D 3D modeling to ArcPro multipatch. The story map explains the process undergone and what the next steps are in comparing Pix4D are ArcGIS Pro's multipatch.

https://storymaps.arcgis.com/stories/e801e9e86d444b2487fdbe872f80b9f5