User testing - list of users (Dec 9 - 9-12)

collated dev tasks

November 19th

metadata:

site title: “product name TBD” (get user input)

logo: ncstate + usda logo

favicon: currently nc state logo, can be replaced with a simple drone

add description/caption to the video to show the user that the map is on the right? but we already use an “arrow” in the text above the video to indicate that

November 5th

• devs ready for minor ui changes

• ci/cd - dev prod + dev local done. prod left

• superusers to use app 1 on separate VM

• @jinam talk to jevon - get srs server on campus

• all superintendents attending, some breeders - total 10 so far

• akshat - code for UI changes done

• ol existing tools for alleyways/manipulation of plots + does mapbox have anything?

  • dropdown to add alleyway height/width - initial

  • + allow user to edit it

October 15th

• Dec 9 - 9-12 in-person user testing

• setup new VM with OIT for app 1 - on-prem for superusers

• OIT - will in the future create new VLAN for research computing

• jinam - debugging issues in HPC processing

October 8th

• chris submitting ticket to cals_it to install app 1 on the lenovo server for ‘superusers’. Jinam work on setting upload timeout to many hours to allow for the uploads.

• the December 9th week is #1 choice for user testing with GameTheory. User testing will be a face to face 2-3 hour workshop like event. Fikret has a graduate student class that week that involves folks like Joe and Jeff. Chris check with Fikret for schedule and possible cancellation due to low registration.

• setting up development environment for app 2 on the same VM. Will use containers to separate development and production environments. Jinam working on it.

• Continued bugs on sfm code that prevents use of the gpu’s and subsequent timeouts on the cpu’s. Possible use of Vyshnavi given all the competing demands on Jinam. Chris checking with Mikah and Steven about using her time.

• GameTheory coming into town for face to face user testing.

•  

September 24th

• Question: what is the resolution of images for each sensor

• GT - Jinam gave overview, Susan getting unityID access

• Research station - geojson files for field names

• Dev env for app 2 - scope of the development? app 1 + 1.5 + 2 or app 2 in a silo

• CI/CD or separate prod and dev env → @Jinam next step

• Edit puppet files → stop setenforce 0 + nginx default config for app2

• Request service account for app1.5 → Jacob talking to Eric

• One more flight for central and sandhills in the next 2-3 weeks

• Chris - move metashape + pdata + raw folders from sandhills locker to a “free-ish” locker

September 17th

• separate scripts for moving Sandhills 2023 data into the MongoDB. Jinam has already started script. Not a high priority.

• Jacob - try using multiple GPUs + using whole node

  • long term looking at what in ODM will not run on H100’s

  • We are using whole node because LSF was using lots of cores and causing concern. LSF requests

  • Chris getting GameTheory onto Unity ID with Susan

  •  

  •  

•  

Sep 10th

• Rob - processing of data

• User testing

• delete flights? - delete processed data or delete the raw images as well. interface to manage flight data - database + files.

• get display name from the pilot

• sat img for spatial query / use mapbox?

• use Rob’s geotiff of field names to overlay on sat img

• tooltips etc for better ux

Sep 3rd

• Amanda - ontology stuff update

• Catchup on adding data for 2024 - running through the whole pipeline

• Jacob - talk to HPC team about preferred queue

• NoSQL db - dashboarding + other insights/saved queries tool

• Cloudiness dropdown - code done, need to merge

• Jinam - update HPC codebase (add cleaned up code)

• Jinam - working on adding research stations to DPP

• Additional data product - download one image for the whole grid too along with images for each plot

• User testing in October

• Plant breeders or crop production

  • Anna locke

  • Ben

  • Vasu

  • susana

  • ruben alvarez

  • matt krakowsky

  • rachel -

  • dominic

  • jeff dunne

  • nonoy bandillo

  • jim holland

  • reza

  • craig

  • hamid

  • carlos

  • jessica spencer

• superintendents

  • phillip winslow

  • jeremy martin

  • john Garner

  • keith starke

• Superusers

  • jeremy davis -

  • jason ward

  • brynna

• Involve GameTheory for user testing

• UI updates: Tooltips for users - shift + cursor for moving map, etc/ feedback+home button on last page

• Eric Sills question for Chris. GitHub actions - service account idea.

Aug 27th

• Jevon - going to sandhills to figure out networking problems by Friday

• Jinam - change orthophoto resolution (parameter updates for singularity)

• Multi folder upload - does it do recursive traversal?

• Rob - flying over sandhills and central

• Both P1 sensors - error - stops taking pictures midflight

• Col with and row width + grid should get you 80% of the way as far as accuracy of the plots is concerned

• Need ability to add multiple grids on the same map

• Next steps discussion on version 1 and 2

  • Joe- just delivering plot images

  • Rob- threshold masking should be in scope but not ai enabled. Use HUE slider or something similar to mask out soil.

• App 2 running on vmfarm - http://dronepilotapp.psi.ncsu.edu/

• Next dev steps for scoping

  • Alley way stuff

  • Slider for index/rgb values to create a mask

• Changing the display name (maybe can change to date and research station with a smaller component with some other info)

• Change spatial query map to satellite imagery

Aug 20th

• Ask users in app 2 - details about breedbase details required - trialDBID, studyDBID

• Pop-up for user authentication of breedbase/BrAPI endpoint

• Does it make sense to have a separate page for exports - create a decision tree to direct users through different flows?

• QR code for trial details has info in breedbase - https://ncsu-maizebase.net/user/login?goto_url=%2Fbreeders%2Ftrial_phenotyping%3Ftrial_id%3D207

• Get ontology info from imagebreed - https://github.com/nickmorales/imagebreed/tree/master/ontology

• Get Study Name from the user in info modal

• Jinam - taking over app1.5 problems from OIT + working on script to import previous data into db

• Akshat - working on multiple folder upload for app1

Aug 13th

• HPC - code problems fixed, ortho intelligence moved to singularity

• HPC - db networking problem still exists

• Using azure db as the prod db for now. Will move to dronepilotdb when networking fixed

• Chris talking to Andy at 1 about storage locker creation

• Better UX - allowing uploads of multiple folders

• Cloudiness - add dropdown instead of text - fully sunny, fully cloudy, 0-20, 20-40, 40-60, 60-80, 80-100% cloudiness

Aug 6th

• Limitation - app1 only able to upload a single file, requires user to select a single folder

• App1.5 networking issue should be resolved by Gregory, Jinam will test it today

• Add last year’s processed orthos to app2 - direct add into the database

• Adding folders for specific users + multiple locations on OIT (recheck next week - Andy said would be done by Tuesday)

• Plan to work on adding authentication

• App2 - we have schema for exporting, can interact with drawing plot boundaries

• Next step → reconfigure app1.5 and app2 to deal with HPC structure

• Manually will add the already processed (2023) data to App2 - Jinam

• Maybe have Akshat test the BrAPI export in sending data to ImageBreed endpoint

July 30th

• Jinam + jacob meeting for app1.5 done. Jinam updated code for Jacob’s hpc processing scripts

• Jacob - to do unit testing + pipeline testing with small dataset by end of week

• Jinam - make test database entry for small dataset used by Jacob

• cron will run at midnight everyday

• Jinam - working on adding dropdown + maintaining folder structure for separate folders for research stations

• code changes to app1.5 will be required when folders are separated

• After testing of app1.5, Rob’s uploaded data will be processed

July 16

• app 1.5 is still not running Jinam is meeting with Jacob July 30th

• Joe tested installation of app 1 and it wasn’t bad. Still concerns that the 20-30 people using drones might struggle with the installation. Rob and Joe more experienced with software.

• Looking at installing app 1 on a server on campus. Either the VM farm or a stray server in CALS. Cannot be on the Sandhills workstation since it will soon be at the station. Chris following up with Jevon.

• Follow up from meeting. Chris and Jinam met with Andy Kurth from research storage. We are going ahead with creating folders for ALL stations and one to store the miscellany of super users.

July 9th

• Rob has installed app locally on his computer. Was relatively simple.

• Needed Git to install front and back

• nginx installed

• exif tools

• node.js

• Chris will find a server to install on for the Raleigh based researcher that doesn’t fit our model of stations first.

• Amanda making a generic poster of drone project for asa

• Brynna and Rob doing a turf specific poster at asa that also talks about drone software

June 11th

1. question for a user like Rob and Brynna. Could we bypass app 1 and use Globus? Is anything fundamental happening in app 1 like rewriting file names. How would we get the metadata in there. Answer: Use app1 installed locally.

2. Intentions with Phillip. Does he have the updated network? Even without the updated network, can we install App1. Ask Jevon about how to install there if no linux server in the closet. Add some storage.

3. Timeline for testing.

4. Update from Jacob. Jobs can be submitted now from DTN. Only piece is missing is th

June 4th

1. cron - check flights from yesterday and check status of it (db read) - check num_files

2. flight(s)*sensor (UUID) needs to be processed

   1. each one is submitted as their own "job"

   2. job title should be same as flight*sensor

3. flight a(sensor1) - generate_ortho

   1. status = processing (db update)

   2. hpc processing for ortho generation

   3. generate cog using rio

   4. hpc processing generate veg index files else docker is ready

   5. status = processed (db update)

   6. status = processed, export ortho/dsm/pt cld

Action Items:

• Jinam and Rob will examine a few of the example outputs produced by Jacob to insure we are happy with the outputs and then we will start doing this in masse.
  outputs at W:\transfer\benchmark\HPC\gpu-a100\code\images

• What gpu is in the Sandhills machine? Chris asking Jevon

• Do we need to buy a node with Eric to get at the top of the queue?

• Run the exact same benchmark on the sandhills that Jacob has run on HPC.

Open

• Loading multiple folders - app1?
  - drag and drop a folder does not work - opens new tab

May 28th

• Gregory’s team can no longer make VM’s. VM1(MongoDB) they still could. Gregory just got permission finally to make VM2.

• DTN cannot access the database. Jacob working on opening a port or reverse proxy to make it happen. Or we just go back to the file method with a cron job checking a file every hour or so. Its a bit of a hack so hoping database works. Chris checking with Andrew about other triggering options.

• Purdue plot- alley system. Jinam get the software up and running.

• Export idea - Brapi compliant talk to Akshat

May 21

• Data transfer node (DTN) special policy in the firewall for large data transfers.

• Triggers: in theory globus flows could serve as the trigger.

• Cron job has to watch the research storage locker for new files that have not been stitched.

  • number of files, status of the job flag

  • flags: data uploaded, image transfer complete, processing, processed

    • App 1 which is a local app will count files, checksum and write to database.

    • DTN cannot communicate directly with the database. Hence, app1 will also write the metadata to a file in OIT storage. DTN will access OIT storage to get metadata (file count, checksum, etc) and check the folder for the flight to verify. Once the data transfer is completed, orthomosaic processing will be triggered.

• Jacob/Jinam - also look into trying to access azure db server using DTN

• Chris

May 14th

• App 1 available for testing at https://srs-uav.cals.ncsu.edu/ This is running on the Sandhills workstation. NOTE: you have to be on campus or on the vpn for it to work.

• Rob playing around with web mapping - http://152.7.196.7/cc/

• Collected images at Central Crops on 8th - Planning Sandhills for this Thursday(16th)

• Discussing possibility of using Purdue ‘Data for Science’ on App2

  • https://github.com/gdslab/data-to-science

• Can we elevate our status in the queue for certain nodes.

May 7th

• Chris checking on clayton IT infrastructure. Question about where app 1 api lives. On a linux server in the closet? Or on a windows desktop? For now, using Sandhills workstation while its still in Raleigh. This is a problem just for Rob’s group at this point. Chris checking with Jevon on this plan.

• trigger ideas for app 1.5. Add a file count feature to Mongo.

• Cron job on one of our VM’s that checks if all the files are present yet and if so start the stitching pipeline.

• Phillip Winslow asking about app 1 along with others

• VM1 - just mongo DB, cron on login node

• service center

VM1

• Just the Mongo DB

VM2

• app 2

Login node -

• where cron job will live

• cron job runs a mongo query that checks file count for all flights where an ortho has not been created yet

• has to have read access to MongoDB

April 30th

• Help with App1 install on local computers

• Starting data collection at sandhills tomorrow (May 1st)

• Setting up flight plans for Central Crops - prioritizing trials (High, Mid, Low, priority)

• Have Jevon setup Sandhill's workstation in the sandhills

• Gregory - working on database deployment and then will work on app 2 server

• We should probably start documenting the processes involved in OIT communication. It should make it easier for the rest of the team

• Jinam/Jacob - look at app1.5 singularity code cleanup and deployment

• Jinam - work on app 1 multiprocessing while checking and writing files

• Jinam - start exploring github actions for app 2 CI/CD

• Jinam - talk to Mikah about integrating a feedback button for app 2

April 23rd

Purdue has a github open source project for downstream data processing. Beyond what we have done.

https://www.ag2pi.org/workshops-and-activities/field-day-2024-04-24/

Jacob’s experiments. Not sure the comparison with workstation processing times is using the same dataset.

Asking Trevor to set up github actions for VM farm. cc Jinam in slack. Normal Docker for app 2. For app 1.5 the ODM and our scripts that creates the orthomosaics.

Our scripts will only run on the A10 gpu’s right now. Singularity containers will not run on the A100’s and H100’s as of now. They are still being tested.

Rob and Brynna are starting now with field data collection. Wants to install app 1 on a campus computer in their lab.

April 16th

• Jacob - coordinate with Gregory on getting the MongoDB server started

• Jinam - debug exiftool issue for 10K files

• Joe, Rob, Anna have volunteered to be the early testers of the tool

• Set a date for early testing

• Jinam - start working on deployment script of app 2

• CI/CD thoughts - git actions (commercial git) / azure devops (enterprise git)

  • Jinam and Jacob will meet on their own to work on setting up the deployment pipeline.

April 9th

Clayton expansion
Joe has some fields in Clayton that he would like imaged. We need the names circulated soon for other programs. Anna doesn’t have anything at Clayton this year, she has one field at Sandhills. Ben has been in touch about another field they have.

Workstation moving
Jevon is moving the workstation next week.

Serving COG’s locally
Jinam experiencing cog problems but should be overcome soon.

April 2 2024

• DJI just announced ‘DJI Dock 2'

• How many digits needed to store vegetation indices

  • three significant digits in general for NDVI

  • Jinam making executive decision for now

• Open

  

  Proposed Data Processing workflow

  • Store the index layers for simple/quick access

  • calculate the plot-level statistics after user draws plot boundaries - store in mongo with plot-id’s

  •  

Mar 26, 2024

• Finished code for generating veg index values but it has high read and write latency.

• Jinam to explore multithreading at plot level index values (for read latency) and using HPC to generate veg index values for a flight (for write latency).

• Jacob to check HPC processing and getting a service account for HPC.

• GPU behavior seems erratic. Jinam - look at ODM’s usage of GPU. Jacob - look at continuous (graph like) usage of resources when running on HPC.

• Orthophoto generated using ODM to compare CRS with Metashape.

Mar 19, 2024

• Keith Starks - very enthusiastic and looking forward to this work

• Talk with Jeremy Davis (OVT dropping enterprise drone deploy)

  • who could I talk to about setting up web ODM for extension agents to process imagery?

  • talk to Rachel about setting up web ODM on OIT for extension agents to use

• List of veg indices we’re calculating and a list of sensors we’ll support for the indices.

  • (See Jan 30th notes for indices)

• List of zonal stats

  • Red, Green, Blue, Red Edge, and Near IR reflectance, and some common vegetative indices
    (See Jan 30th notes for indices)

• Comparison between ODM and metashape CRS

• Version 1 release should have a media release with it (article) - Dee

• Should we start the process of applying for a exemption that leads to an exemption for drone in a box

  • Tom and Evan (NCDOT - ITRE)

Mar 12, 2024

• Meeting with Keith Starks (Central Crops) this Friday (last Friday was cancelled) to discuss plans

  • possibly flying Strawberry field for Jing Zhang and Gina Fernandez

• How do we have others access the data on OIT storage via GLOBUS (possible demo for Rob?)

• ODM/Metashape Comparison

• DJI Care Renewal for P1 ($750) and Matrise 300 ($1200): $3900 Total

  • https://store.dji.com/product/dji-care-enterprise-basic-m300-rtk?vid=102921

• How do we calculate the different vegetation indices?

  • https://towardsdatascience.com/segment-satellite-imagery-using-ndvi-and-rasterio-6dcae02a044b - veg indices

  • https://pythonhosted.org/rasterstats/manual.html#zonal-statistics - zonal statistics

• Demo: plot ordering and plot numbering

• Discussion around data products to export - csv of plots, plot images, orthos, etc.

• Domain Names? Should be made according to a new product name (instead of “drone pilot”)

• Thoughts around Tile server/ ways to serve cogs (will need URLs instead direct blob files)- we solved this right?

• Send Jinam date that contains visible GCP for comparisons

Mar 5, 2024 | IAM discussion: Drone Pilot Project

Attendees: Jinam Shah Jacob Fosso Tande Chris Reberg-Horton Billy Beaudoin

Interfacing NCSU IAM system with Apps being developed

Notes

• Identity need to access data on the oit storage. Before that happens, the researcher has to authenticate, this is on App2

• Researchers are importing their plot plan and doing the editing. There is need to protect the editing that happens.

• Application is needed to load data from the instrument. To use application 1, user must have a unity ID.

• Stack: react frontend and backend is python

  • Do a reverse proxy on the frontend

  • NCSU servers uses Apache and not nginx, for nginx, reverse proxy will be needed to access shibboleth

    • https://github.ncsu.edu/OIT-IWS/ncsu-shib-sp-proxy - Example of doing a Shib SP

  • Be great if Billy can provide documentation on how to do the implementation.

  • https://docs.shib.ncsu.edu/  

  • Request form: https://go.ncsu.edu/shib-access-request  

Action items:

March 5th

1. PRIORtITY 1: Test run image processing on HPC (ODM singularity instance) on full set of image
   - compute time, start determining resources needed

2. PRIOROITY 2: Plot order and plot labeling

3. Install web server on sandhills work station (already installed)
   - Rob may install tile server and play around some

4. Timeline for moving workstation to Sandhills - Second week of April (waiting on VPN stuff)

5.  

February 27th

1. Web server on VM - nginx

2. Get Jacob all the software needed for the VM

   1. nginx

   2. mongo DB - community edition (set up as production)

3. Central Crops - see if they already have desktop for UAV (Jevon)
   - not a priority as we’ll (Rob & Brynna) will be doing the flying and can upload from NCSU campus

4. Talk to Keith Stark about flying at Central Crops (Chris will reach out)

5. Hold back on figuring out GPU + multispectral with ODM issues

6. Uploading Plots plans

   1. discussed with Chris/Jinam

7. Get Toughbook/Fieldbook from Slater

February 20th

1. Determine the VM size (memory, disk size) , configuration (ports to open, static ip, network, os to use)

   1. How long will this VM will stand (likely duration of the project)

   2. Can the existing configuration management be used

2. Figure out why Jinam can't use the GPU with ODM on the sandhills workstation (CUDA)

3. Priority 1: ODM processing with GPU + multispectral image processing when some are corrupt

4. Priority 2: Enable spatial query on mongodb/python for varying CRS output from ODM

5. @jinam: connect with Jevon and mount OIT storage to sandhills

6. @jinam: convert past TIFFs to COGs

7. Keep Sandhills workstation at PSB until first beta testing is complete

February 13th

1. Can people see other peoples plots? Yes. They won’t know the treatments etc. because that info lies with the researcher.

2. Authentication? Not necessary to filter for a PI. But would be necessary to protect plot layouts which are a time investment for the researcher to input. Jacob will approach the OIT group involved in identity management.

3. Study level metadata. Chris taking the lead on traits.

4. Visualization of the orthos and the ability to query them. Priority #1

5. Ability to download the ortho from what they defined as their polygon (study in BrAPI speak). Names will be BrAPI compliant. Version 1. Version 2 maybe ability to push data directly into a breeding database? Still thinking.

6. Get all of the plot tool components working.

   1. All rectangular plots shapes possible. Dimensions of each plots, allies, etc..

   2. alleyway dimensions

   3. number of plots

   4. Labeling: Serpentine vs dead head etc. Copying logic from Juniper systems. Esleyther has a ‘Toughbook’ with it installed.

   5. ability to modify an individual plot, not just the whole study. This is to accommodate oopsies!

7. Zonal stats (summary of pixel values based on plot polygons) based on user defined plot boundaries
   https://pythonhosted.org/rasterstats/
   
   

8. Associate and display stats (e.g. veg indices) at the plot level

February 6

• add ability to get single image per plot

• export whole ortho

• add numbers

screen show plot layout

include number of rows, number of columns, plot width, plot length in feet or meters.

• Other metadata for ‘study’

• traits taken for study

how are we getting dtm?

Jeff: plot level analysis as a way to inform him of when to go and sample, plot level value is mainly what they want. Use ocassional picture for a presentation. Raw images could be for training a ML model.

add detail to trait data as a string

no one on campus that Susan is aware of is archiving data at scale

need more names, PI’s program, grab unity ID as they interact with app 2

We could ask station staff to fill out an end of year thing asking what crop, what PI? Additional feature to app 2 for a superuser to add in metadata.

Add in field numbers.

January 30th Stars below mean version 1.

• For mockup- possible vegetative indices

  • Common RGB indices

    • GRVI (Green Red Vegetation Index)

    • VARI (Visible Atmospheric Resistance Index)

    • GLI (Green Leaf Index)

    • TGI (Triangular Greenness Index

  • Multispectral Vegetation Indices

    • NDVI (Normalized Difference Vegetation Index including red-edge)

    • NDRE (Normalized Difference Red Edge)

    • SAVI/MSAVI (Modified Soil Adjusted Vegetation Index)

    • EVI (Enhanced Vegetation Index)

    • LAI (Leaf Area Index)

    • NDWI (Water Index)

    • CHI (Chlorophyll Index)

  • Study Metadata

    • Crop type/Species

    • Planting Date

  • Traits collected in study

    • yield, insect damage, quality data

    • Biotic/abiotic stress ratings; phenological; yield

January 23rd

• Debrief of 2024 Superintendents meeting - general interest, concerns, thoughts on including new stations this upcoming season

  • current capabilities vs equipping additional station(s) and training staff?

• What is the interest in a Hiphen visit

  • http://www.hiphen-plant.com

  • February 19-23

• Additional Tower/workstation (check with Devon)

• Targeted and Non-targeted cost estimates for deploying on station

• Meeting with Jeremy on Sandhills (Rob/Chris/Amanda) within next week

• Move forward option of drone in the box (start thinking about funding)

  • Concerns over DJI and FAA requirements

• Expand locally at Sandhills with existing technology, expand 2024 activities to additional station

• Where to expand for the next station?

  • discussion on need for money/resources to do some of this

• Philip Winslow (just needs an RTK drone) - could run both the stations in Kinston

January 16th

• (Pushed to next meeting since folks weren’t available) Start conversation on plans (flights) for this upcoming year (sites and equipment)

• Mostly Jinam answering Brian’s catch-up questions on the background of the project and the architecture, where each file lives at what step, how we’ll handle authentication, etc.

  • Authentication: Current plan is that only App 2 gets auth, since App 1 is served from the local research station. Does that mean that we’re essentially setting up user groups where everyone from Research Station X is a viewer of the same data? Should there be narrower scoping to assign each image on a per-user/per-group basis? We decided that any auth decisions shouldn’t happen until after the Feb demo, since we’d rather have more functional features to show then and get feedback.

• Next step we identified: make sure that serving tiles from Globus via URL is fast enough for App 2 front end.

• Followup items: Brian will make sure Alyssa knows to get a tablet with Fieldbook to Jinam.

January 9th

• The front end is now sensor agnostic by just choosing a very high resolution for the configuration file and letting ODM go from there. As long as someone doesn’t fly crazy low we are ok.

• Jinam has started on how to define plot overlays on the images. The rotation of the image is already implemented. We will use ‘Fieldbook’ terminology in version one to deal with concepts of row column layout of plots. The ImageBreed videos are also a resource.

  • Rob recommends against rotating the map, but rather rotate the plots based on the users interaction with the map.

• We debated how to assign plot numbers to the row, column designations for each plot. Decided that feature will be done by the user via file import? Or is that still being debated?

  • Rob thinks they should be stored internally as row-column but provide the users some options for naming plots. (row/column versus serpentine, different start number (e.g 1001) etc..

• Waiting for more feedback from Globus on whether our application can access the unique URL’s for every image. If so, we will not need a tile server and so our locally installed app can function how COG’s function in the cloud. If that effort fails, we will just insert a tile server into the application.

Jan 2nd, 2024

• PRIORITY 1 (App 1):
  - Fix https upload issue
  - Add functionality to App1 to upload multiple folders

• After upload, compress pictures (lossless) into a single .tar file

  • transfer to campus via globus

  • Decompress and store RAW on central storage in decompressed state

  • Stitching into the orthomosaic

  • Store all data products uncompressed

• Should we store raw images in an ungeoreferenced way or after georeference do we store only that. This question can be answered after launch of version 1. The only reason to ask this question is to conserve storage space. If we store both its double the space. Check in with Francisco for his opinion at some point.
  - https://community.opendronemap.org/t/extracting-orthorectified-images/1372/3

• PRIORITY 2 (App 2)
  - work on drawing COGs using Globus generated URLs
  - setup App2 on webserver for demonstration purposes

• Send out doddle pool for larger group meeting. Chris sending by tomorrow at latest.

December 12th, 2023

On App 1: Force the pilot to separate the uploads where 1 upload is data collected from one sensor. When flying multiple sensors, one flight would require two separate uploads. Consider changing the mongo structure from ‘flight’ to 'upload'

• PRIORITY: Work on converting existing Geotiffs to COGs and having one display in a web browser using either leaflet or OpenLayers
  PRIORITY: Look into JavaScript based vector drawing 'plugins' for leaflet and/or openlayers
  Leaflet Draw API: https://leaflet.github.io/Leaflet.draw/docs/leaflet-draw-latest.html#l-draw-toolbar
  Possibilitys:
  - https://leaflet.github.io/Leaflet.draw/docs/leaflet-draw-latest.html
  - https://github.com/geoman-io/leaflet-geoman

December 5th, 2023

• PRIORITY 1: Setup App1 on Sandhills workstations
  → setup landing page where images files are uploaded, get upload working with db backend, have metadata written to db, and get Globus transfer to trigger after image upload

• PRORITY 2: figure out reasons for ODM application failure with high res images/compression settings

• Web server installed on sandhills workstations. Setup port 80, Jevon will do this..

• Setup a VM on the VM farm (VCL? - is it stable enough)
  - short-term: setup on sandhills workstation and get up and running
  - long-term: App 2 lives on VM farm
  → decided this will be setup when we are ready for production server (hold now)
  - COGs on central storage or VM storage
  
  
  

November 28th, 2023

• Deploy app 1 on research station workstation and get working

• Setup VM1 as ('Manager VM' Cron job, Mongo DB) on VM farm

• VM2 will host application 2

• Set up singularity/docker instance on sandhills workstation to get processing pipeline up and running (testing)

  • setup same structure on SH workstation (VM, singularity, etc...) for ease of migration to HPC and VM

  • decide on singularity/docker for SH workstation (not a overly big decision)

• Run an ODM process on HPC for initial resources testing (work with Jacob on setting up on HPC )

• ask Jevon open up port 80 on SH Workstations

• Meeting with Micasense on Friday to figure out Altum-PT skyport connection issues
  → figured out the issue (2TB CF Express card was incompatible with Altum-PT)

November 21th, 2023

• Discussion of backend server needed to serve UAV images/orthos (App 2) or not?

  • Maptiler as backend tile server vs TileServer vs TiTiler vs Terracotta

  • right now, experiment with severing COGs directly with OpenLayers

  • https://github.com/cogeotiff/rio-cogeo

• Discussion on frontend Leaflet vs OpenLayers (App2)

  • leaning toward OpenLayers, larger codebase, but lots of existing plugins

  • need to look into drawing plugins

November 14th, 2023

November 7th, 2023

• A config file will store the sensor parameters outside EXIF for obtaining our desired final resolution.

• Exploring Mongo as the DB. Yes it stores geographic data. What are the advantages of using Mongo vs using the ‘JSON’ field in PostGres

• API that will serve the querying functions in app 2. Will also handle the write back into the DB of plot coordinates and plot names etc.

• Still working on the calibration methods, not a huge priority for now. Get the rest of the pipeline working and go back to what the calibration equations should be.

Oct 24, 2023

• resolution still not matching. That is just settings

• Make the compression algorithm the same

• Color checking not really there. Time to just do this with OpenCV

• Jinam has mentioned to Rob the EXIF file containing the metadata is not the same for Altum PT that we have. So how do we perform the color calibration?

• Where,

• r is the distance of the pixel (x,y) from the vignette center, in pixels

  • (x,y) is the coordinate of the pixel being corrected

  • k is the correction factor by which the raw pixel value should be divided to correct for vignette

  • I(x,y) is the original intensity of pixel at x,y

  • I_corrected(x,y) is the corrected intensity of pixel at x,y

• In the radiance calculation above, V(x,y) is equal to 1/k

• Is the vignetting model always the same for each sensor or is it for each image? Are there additional tags beside exif we are missing? XMP?

• For now we will assume first two images and last two are the calibration checker. We will have to build a color checker auto detector for this sometime soon. Could be just one image or multiple that has a checker in it.

OOctober 17th, 2023

• Most of the data has been processed manually through metashape for the year

• Automating that process with Open Drone Map continues to go well

• We have about 12 TB of data sitting in our locker in Research Storage

• Jacob helped us manually move all of the MetaShape files to the locker today (another 10TB). We will keep those there for a few months but will likely move or discard in a few months based on how our new pipeline is working.

• Jinam is putting his code into a couple of GitHub repos this week so that we can start porting more of the pipeline pieces onto the HPC cluster

• Rob is starting to share some of the data products with users via Globus. While not our long term way of sharing, we want to show value now.

Sep 26, 2023

• How would project members be given access to the collected data?

  • Add members to the OIT storage for access to all the data?

  • Use Globus collection to grant access to specific data in a specific folder?

    • Project members actively moving data will have access to the OIT Storage.

    • Everyone else will be given access to the data through globus API

    • APP2 once developed should give users access to data.

September 19, 2023

Switching to using Azure as our development environment. Rob and Brynna using the workstation too much to use it for development for Jinam’s work.

Can we put a globus endpoint in Azure? Should be possible, asking Jacob to work on it.

Maybe instead of trying to divorce breedbase from the image part we can just cherry pick the bits we need for app2. This will be much faster. The separation of breedbase will have to be another team’s task. The database code is all written in Perl. The plot/alley bits are all python and should be easily followed.

BrAPI compliance at the export layer in app2.

Sep 12, 2023

For processing the calibration images, which include the reflectance card, we need to autodetect which files contain the calibration cards. No metadata coming from operator SHOULD be used in determining which files. We SHOULD examine first five and last files in every flight for every band to identify these calibration images. If ODM does not contain this functionality we will find a path with opencv or other libraries.

Aug 22, 2023

To include in design document. What data products will be computed by default as part of pipeline. And which data products would be on demand.

Do we have an internal meeting to create the design document as our ‘Hackathon’? Or just invite Francisco? We just go forwarded with our own scheduling and then add some from Cornell but not necessary for them to attend. Francisco is the most useful

Chris circulating design spec documents for us.

August 15th

• Jinam will start exploring open drone map sooner than expected since MetaShape floating license is proving slow.

• We plan on assisting Cornell with separating ImageBreed and BreedBase and getting ImageBreed away from Perl.

Microservices envisioned

• 1. Stitching

• 2. Web front end to draw in where your plot boundaries. MapBox and javascript front end

• 3. Plot level data products produced. Python/R backend that generates plot-level values/statistics based on the locations of plots within the georeferenced plot map (from service 2) and the outputs from the #1 (orthos, dsm, and point clouds)

• 4. Data analysis portal - probably not or later version. Could install Jupyter Hub and create an analysis portal but overall we don’t think this is the right crowd.

August 1st

• Charging stations arrived and we will take down to Sandhills tomorrow

• Sandhills flights planned for 8/2 - looks like a nice day, will plan to fly both Multispec/thermal and RGB

• Metashape did not recognize the GPU on sandhills workstation. Working with Jevan, Jacob, Jinam on getting drivers installed and Metashape setup to use GPU - hoping will reduce processing time significantly (Complete - Jevon)

• Needed to change mount point from /mnt/data/raw to /mnt/data

  • This created a need to restructure the folders on the sandhills workstation as seen below
    
    

• Transferring last two weeks of sandhills imagery to workstation today - will start to process as soon as GPU is setup and ready

• Timeline for moving workstation to sandhills?

• We should start to think about metadata. A good place to start would be to make a two column table. The first column could be metadata extracted from the images and column two would be what we want to see as metadata.

July 25

• Continue to collect Sandhills imagery weekly and transferring to sandhills workstation

  • When weather permits, we collect both RGB and MS/Thermal in high priority areas

  • Processing manually, about 1 week behind

  • Processing on workstation can still take 6-8 hours on larger missions. We typically queue up 3 flights and run overnight, and most mornings the processing is finished, but not always (typically ~8 missions collected in a day)

  • We should look into the cluster based processing capability of metashape

    • https://jenacopterlabs.de/?p=862

    • https://www.agisoft.com/forum/index.php?topic=10459.0

• Still waiting on charging stations and Altum-PT sensors (asked Benchmark for an update yesterday, have yet heard back)

  • Currently charging extra batteries in Raleigh and switching out with Jeremy in Sandhills

• Jevon is helping change the 20TB mount point of the sandhills workstation from /mnt/data/raw to /mnt/data

  • We needed space for ./metashape, ./pdata (processed data), and ./scripts to sit at the ./data level

  • May need to do some reconfig with Globus after moving mount point

  • Once setup, we’ll start moving and processing data again

• Purchased additional SD and CFExpress cards for sensors. Altum-PT collects A LOT of images, 4 flights is ~180GB of images

• Effort to Automate Metashape pipeline (Metashape Hackathon?)

  • https://github.com/nirwan1265/metashape

  • https://github.com/ucdavis/metashape

June 27th

• Plan to fly this Thursday - looks like nice weather and hope to collect multispectral data

• Setup M300s with new KML flight plans - new flight plans are designed based on priority (will discuss in detail at next meeting)

  • Developing map that will illustrate missions and priorities

• Still waiting for battery charging stations - hopefully next week

• Globus setup last week - will transfer raw and processed data this weekend

• Jacob Fosso Tandecould you setup a ./pdata/ folder on the sandhills workstation for the processed imagery (as well as give proper permissions for transfer)

  • /sandhills/raw and /sandhills/process/

  • You should set up the same file structure at source then copy sandhills and everything will map correctly.

  • I do not have access to your source system but globus script that can do the setup with very little interaction. After the first run, the transfer will happen automatically. However , I do have another script that you can use anytime provided you give it the correct input.

  • If this needs a separate globus collection, could you set this up as well

    • No, a collection for a station. Beneath the root folder, other folders can be added as needed

• Brynna will post a processing report from one of the larger metashape projects

• Jinam is still hung up with his visa. Susan working with him closely

June 20th

• Poor weather - perhaps Friday Flight

• M300 setup and transferred to Sandhills

• Setup standardized flight missions and will copy to Sandhills Drone

• Working on processing UAV imagery

  • Ran out of memory on local machine running 3 missions

  •  

  • Charging stations have yet to arrive

• Tested SFTP transfer from williams to PSI on for single date - stopped @ 6+ hours and about 50% transferred

  • Jacob will show us how to setup Globus later today and we will test

Thoughts:

Do we need to set up a tracking sheet to record flight dates? Google sheet?

Where do we stand on metadata collection?

June 13th

• UAV equipment setup and registered with FAA

  • Charging stations have yet to arrive (hopefully later this week)

• Driving down UAV equipment to Sandhills today

• Collecting data at Sandhills today

  • Training Brynna on M300

  • Flying both UAVs with both pilots

  • Will join today’s meeting remotely

• We will start transferring imagery to Sandhills workstation ASAP

  • Need the IP address and login credentials (unity?)

June 6th

• Brynna starting work today - will start training her ASAP

• Drone equipment has arrived!!

  • Multispectral sensor still 8-10 weeks out

  • Charging station 2-3 weeks out, but we have backups (thanks Jeff D!)

  • Will take a few days to get setup, registered, and firmware updated - should be ready to fly next week

• Planning Wednesday or Thursday Trip to Sandhills for image collection

• Will start pushing images over to workstation as soon as everything is setup

  • Also plan to manually process until front pipe is setup

• Globus configuration:

  • /data/raw : Path accessible to globus automated transfer

    • Data is automatically transferred every day at 12:00am

    • Data differing in checksum is transferred

    • Preserved timestamp at source

  • Should the shared directory be owned by a service account?

May 30, 2023

/rs1/shares/cals-research-stations/sandhills/raw read only, even for developers

/rs1/shares/cals-research-stations/sandhills/processed read only for users, write access for developers only

Ask Jevon to setup accounts on workstation

May 23, 2023

ImageBreed was built to serve Kelly’s lab. Was built on top of BreedBase.

They use Pix4D, but there is an open source that can be used but is slower.

Have to store your trial info in BreedBase.

Images get pushed into BreedBase.

Some design choices that need to be fixed. ImageBreed changes were not pushed back to the BreedBase repo.

Want to take the image analysis pieces and make it a stand alone application.

May 16th

First Flights Conducted May 11th

4 flights - Zenmuse P1 sensor

Blocks D, AB, E, and F (flown @200ft)

97 GB of images (5700 RGB images)

Can fly 30 minutes at 200ft with P1 sensor (0.30in/px - 0.7cm/px) ~40 ac/flight

• 40 minutes to transfer from SD card to local storage

  • Order extra SD cards

• Need to decide on storage space

  • Chris will have Andy setup folder

• Wednesday meeting - agenda/thoughts?

• Schedule interview with Jinam Shah

• Planting Update (Jeremy)

• Setup Github repository (Chris’s team will handle setup - public account)

May 9th

• additional batteries are ordered. All told, each UAV will have 8 available batteries (4 flights)

• Cloud City Drones was required to cancel their bid, bid issued to Benchmark (local vendor) May 8th. We were told 2-3 weeks for DJI equipment (platform and P1 sensor), 8-10 weeks for Multispectral sensor. (They have been strung along about delivery dates for the Altum-PT sensor - see Joe Gauge)

  • Need to

• Update from Joe about flying multiple sensors

• Prioritizing Flights

  • ‘Non-objective flights

    • Goal is weekly flights

  • ‘Objective’ Flights

    • best to identify crops and objectives as clear as possible before last minute (scheduling)

  • Cotton (A3/A5/D2B)

  • Soybeans (A4)

• Prioritizing Sensors

  • If conditions are not conducive to useful data, is there any value in flying the sensor?

    • On partly cloudy days, you cannot ‘easily’ radiometrically correct for shadows within the flights and data - as such, is there any reason to fly the sensor in those conditions?

• First Flights Planned for Thursday (May 11th)

  • Update from Jeremy on Planting

  • Matrise 300 with 8 batteries

    • 4 flights possible

      • Need to prioritize areas and sensor (s)

      • Value of multispectral with little-to-no plant material?

https://www.ag2pi.org/

https://phenome-force.github.io/PhenomeForce/

May16th

Developing plot boundaries - when/responsibilities

May 2nd

Kim has started working for Rob this week (Pilot #2)

Lucidchart link broken

Get Update from on Jeremy about current planting schedule

• Workstation update

  • Workstation will arrive next week. Looking to complete installation of network rack the week of May 15th-19th.

  • Let’s target our initial flights for 2 weeks after planting.

• Check with Chris on Account to use for equipment, travel, etc..

  • Please go through Susan Wassmer for all purchases. We have 5 different accounts actively supporting this effort and the interplay with fiscal years is complicated.

• How to handle initial flights until equipment arrives

  • PO received by Cloud City Drones May 2nd

• Can only fly one sensor at at time (must prioritize flights)

  • Joe is experimenting with both sensors on the drone but uncertain if it will work.

  • Single Matrice 300 available (Alex Woodly’s)

• Autum from Alex, P1 from Joe

• Batteries. We have enough to borrow from Joe and Alex but need to double our current battery order. Doesn’t have to be through cloud city, can be done on a PCard.

• Should we fly non-RTK drones (need to consider GCPs)
  - We are really trying to just fly RTK and we think we have it through Joe and Alex.

• Dividing out two groups of flights. Fields with a clear objective to be flown at 150 ft, and those without an a priori objective get flown at 250 ft.

• Without objective flights should be at least weekly.

• An objective flight like “stress flights” might even be daily? Decision of when to fly

April 25th
Drone purchase will be from Cloud City Drones ($79,807k)
- comes with 1 year enterprise shield warranty

- drone and P1 sensor immediately available, multspec 2-3 weeks out - hopefully

Rob Join Developer meetings when needed or by invite

Hard pass on UAV-pilot from durham-tech

April 18

Meeting to discuss flight plans in Sandhills

Stand counting

Establishing nutrient research system

Drought stuff - plot level data

Crop growth/monitoring

Jeremy & Denise - created a google form

Usually do get a plot map of most fields

General Updates/ToDo

• Get permission from Alex Woodley to use to match what sensor suites are coming in - Rob

• UAV-pilot from Durham-tech

• More batteries (4 more) + charging stations (3k)

Flight Planning Discussion

• Cotton [RED]

  • D block: D2 D3, D5, D7 (dependent on Todd), maybe A3 (dry cotton)

  • B - late planted cotton test

  • Canopy Temp, NDVI, Canopy Height

  • Planting first week in May

• Soybeans [GREEN]

  • A41 (irr/non irr trial) - thermal (Anna Locke)

    • Perhaps time series during drought

    • Gas exchange focus

    • Can you replace wilting ratings? (again, ML approach opportunity)

    • Planting first week in May (more like single crop)

  • F Block: F4A F4C (Ben Fallen)

    • F2D, F4D - large fields (wilting ratings)

    • Irrigation is cut, ratings start

    • Planting first week in June (needed because of double crop)

  • Ballpark July/August wilting starts (3-4 days without rain)

  • Wilting ratings (perhaps ML opportunity?)

• Peanuts (Jeff Dunne) [ORANGE]

  • A43

  • Planting (maybe) mid May

• Corn (Chad Pool) [BLUE]

  • A42

  • Soil moisture meters and compare to weather station, works with Jason Ward

  • Planting this week

• Wheat [PINK]

  • C Block (harvested in June)

• Turf

  • E2

• Hort Crops - Sweet Potatoes, muscadine, others [YELLOW]

  • B2B/B2F

  • B4A - Peach yield test

• Tobacco/Soy - herbicide trial

  • (B4B)

  • Planting week of May 15, plant both the same day

• Sweet Potatoes

  • B2B2 (plant in June) - not exactly decided on field

Initial Plan (Amanda & Rob)
Blocks D, A, E - specific targeted flights, weekly, varied resolutions based on objectives
(Brynna & Rob)

Need to refine goals and objectives crop specific

Blocks F & B - SH staff handles weekly collections plus one-offs as possible uses-cases emerge

More batteries?

April 11, 2023

- Brynna signed to join Rob’s program as student (should be showing up ~June 1)

- Joe and Rob are going over equipment bids with Susan Wassmer

Looks like equipment will be available 2-3 weeks after they’re received

- Rob will start collecting data and cover until Brynna is on board

Chris - AerPaw company (open source ecosystem) that runs the 5g out at Lake Wheeler interested in us running their drones

Mobile node that runs 5g tower, maintain wireless systems

Wants to know if we can run their drones at some point in the future

Let them know interested in the future but not necessarily right now to alter our purchasing

Might be worth looking at meta-data to make sure our systems work with their schema as well

Jevon - Dell workstation in queue to arrive ($13K) for Sandhills

Ubuntu OS default

Amanda and Joe will ask Cornell to attend a future meeting to discuss their road map and ours.

May 18th, 1-3 as the likely time to meet with the larger group. Kirtan can be here. Need to have him here earlier in the week to prepare.

March 28, 2023

- docker onto the workstation at sandhills

- Stitching - we need someone to explore licensing on HPC’s. Charges based on users or cores could be a problem. Jevon Smith do you have time to look at licensing from Metashape vs Pix4D?
- downstream metadata to add post-ingestion (additional columns)

sandhills or other station, field, crop, researcher

- https://irods.org/ugm2023/

- 30TB available in HPC 'storage’ for storage and processing (central storage)

-> can request more as needed

- drone out for bid (closes Thursday - hoping for Mid-april to May)

Mar 21, 2023

Kirtan available to start mid-May

Set up to meet rest of the project participants at the next meeting 3/28

Discussion on GPS connected equipment available at the stations

What will be used for the particular target projects?

Drone update - going thru purchasing

Workstation update - Jevon submitting request today, probably arrive in 2-3 weeks

Should plan on starting data collection soon as needed then informatics downstream

March 16th Design session

Decisions made for version 1

• Drone operator will load images into a workstation at Sandhills. No metadata will be human input at this stage of data ingestion.

• Plan A will involve moving this raw data via Globus back to the Research Storage on campus where it will land in a storage locker named Sandhills.

• Metadata embedded within the images will be scraped and put into a database and the images file names will be changed to some combination of date and gps location. The metadata embedded in exif files:

  • Data/Time Collected

    • assuming sensor is set correctly

  • Image properties

    • size, resolution, bit depth, ect)

  • Camera parameters

    • Make, model, etc.

  • GPS Location of image

    • Latitude, longitude, altitude

• All raw data will be automatically stitched and processed into data products and written into the Sandhills directory. Subdirectories could be used to delineate types of data products. Note: someone must run some tests on HPC to determine the optimal number of cores, GPU’s, and RAM to run this.

• Data products and raw data will be archived after 12 months. On prem tape or cloud cold storage from the large providers are acceptable.

• Researchers may ‘claim’ data by copying data to their individual lockers.

Web tool

A web based application will allow users to find their data and to compute plot level values. This tool will be relatively simple and it is not meant to be a total analysis and visualization tool. We still anticipate that researchers will do the majority of those tasks themselves within the software of their own choosing. Our app will do the following:

• Allow researchers to search for all data corresponding to a particular set of coordinates and year combination. They can do this via a polygon selection tool that allows them to identify individual fields.

• Allow researchers to upload a field map (with RTK level plot corners) so that plot averages (or percentiles) for any band can be calculated. Plots values should be downloadable. See commercial provider Solvi for indications of intended scope.

• User defines plots in at least 10 polygons

• Additional metadata will be written back to the database based on use of the web tool. Data such as crop, researcher and likely more can be appended to the metadata.

Miscellaneous notes and barriers

• Our database will have to live outside of HPC on the university’s VM farm, which is technically separate from the HPC.

• The web tool will have to run on the university’s VM farm too.

• While HPC is our ‘plan A’, we will use the workstation at Sandhills as our development environment. Work will be moved from that computer to HPC as soon as possible during each iteration.

March 14th

Position Updates: UAV Pilot (Brynna) and Geospatial Developer (Kirtan)

• Looked at Joe’s Metashape performance

  • Jacob suggests - run with/without GPUand test/tune performance

• Can’t work directed off SMB mounts, etc..

• Post-processing options

  • Plan A: Go through HPC (bit understaffed) (production server)

  • Plan B: Sandhills server rack (development server)

    • Strategically involve HPC center, “we can only do X offsite because HPC does not offer Y”

• Discussed pulling metadata from images, storing in DB, query DB for requests surrounding the raw images

Design Systems Buckets

PHASE I: UAV Image Collection, Transfer, and Post-Processing

• Design, develop, and test pipeline from image collection to post-processed data

• Focused effort at Sandhills Research Station (11) - and breeding efforts

• Assess resources and needs (people, hardware, equipment)

PHASE II: Analytics and Visualization of UAV-derived data products

Focus is on developing tools and systems to analyze processed imagery

PHASE III: “Become the daily UAV-imagery provider for all research stations”

March 7th

• Position Update: UAV Pilot

• PSI Drone architectural design session (Thu Mar 16, 2023 3pm – 5pm)

  • Overview of current pipeline

• Research Station Hardware Discussion

  • Adding compute capabilities to the research station hardware (Javon)

    • benefits/limitations/cost?

      • Decided to purchase additional workstation dedicated to this project and housed at research station

        • Rob’s recommendation ~4TB NVME for processing, then perhaps 4-8TB SSD for short-to mid term storage and/or overflow.

        • GeForce 3000 RTK series

  • University/PSI/commercial post-processing compute resources available for post-processing imagery (did not discuss)

    • Not a lot of GPU’s

    • https://hpc.ncsu.edu/About/ComputeResources.php

    • HPC is poorly administered, but should get better, cover all our basses for now with workstation (plan B)

• Flights (targeted versus broad farm-level)

February 28th

• Position Update: Developer Position KIRTAN DESAI

• Position Update: UAV Pilot

  • General Expectations

    • # of flights, how many programs/fields?

      • First year ~1/week at Sandhills

      • Turf, Cotton, and Peanuts?

    • Training and training material development

      • Lower priority at this point, but may be able to work in with some extension efforts

    • Timing-first flights?

• How to incorporate prior work done on ‘leveling data‘ - Rob

  • Bring in design group - leaning toward ‘unstructured’ uploads

• UAV Equipment updates?

  • Do we know the expected delivery for sensors and platform

  • Things to check - with dual mount both sensors RTK capable?

  • Vision for first-year use (1 w/ UAV Position, 1 w/ Jeremy @ Sandhills)

• Hardware Discussion

  • Before any hardware is ordered, need to consider if image post-processing will occur ‘off-site’, at a remote HCP Center, or ‘in-the-cloud’

    • Need to ask Javon about added costs if done remotely (GPU, etc)

    • Perhaps spec out two options

• Arch-Design - need to decide when to bring in Chris’s ‘Geospatial Guy’ (Milad)

  • Design Group Meeting (two weeks)

To Do

• Javon - additional costs for hardware at remote site if off-site post-processing images

• Email susan about Equipment status

• Follow-up with potential UAV Pilot for year 1

• 2 weeks - do the demo (lead off - then talk with design team afterward, 2 hours)

  • Setup meeting to discuss Sebastiano Busato, Daniel Perondi involvement

  • Demo complete workflow with data (could be a long meeting)

  • Include Milad

• Call Amanda about Developer position and if she is okay with the Kirtan (Chris)

To Do (Longer Term)

Setup meeting with Cornell Group to discuss lessons learned, their vision for the ImageBreed Project, collaborative development opportunities

Feb 14, 2023

Here is a dumping of our chats. Will clean up later.

12:05:30 From Chris Reberg-Horton to Everyone:

Notes if you will. First thing to know frequency, flight height, RTK enabled drone so we don't have to have ground control points.

12:10:25 From Chris Reberg-Horton to Everyone:

Need make sure RTK correction is applied to both sensor systems. SD card removed and loaded. Use MetaShap or other similar product where processing occurs on site. On a local server rack. Is this totally automatable? point clouds? Rasterization. return to this topic

12:11:04 From Chris Reberg-Horton to Everyone:

at some point, putting in a plot map overlay to create the plot level outputs.

12:17:42 From Jeremy Martin to Everyone:

Replying to "at some point, putti..."

Who's responsible for onsite processing if server rack located at SRS?

12:18:27 From Chris Reberg-Horton to Everyone:

Replying to "at some point, putti..."

The automated pipeline is responsible

12:32:37 From Chris Reberg-Horton to Everyone:

Globus for data movement. A decision made.

12:49:29 From Chris Reberg-Horton to Everyone:

Version 1 decision. Save all the data. we will study what percentage of the data gets used.

12:59:21 From Chris Reberg-Horton to Everyone:

design to around 100' of elevation for flight