Page Comparison

Attendees

• Daniel

• GeorgiaNathan

• Dan Mashatan from teledynesome autonomy people

• Tong Zhang

• Vibhinn Gautam

• Mihir Gupta

• Director, Auto-leads, PMs only

  • general members not invited, please message us your thoughts in advance

timeline

• only 30 mins

topics to discuss

• explain the team

• explain conops

• can we get an ir camera for task 1

  • 5kg weight limit

  • 940nm, have link for the beacons up

  • spares would be nice for testing

  • early is nice for testing and integration before competition

  • high response around 940nm

    Global shutter

    Resolution of 2MP or greater

    Lightweight (no more than 200g if possible to reduce impact on the drone)

• ideas

  • ask what they think is best

  • https://www.flir.ca/browse/oem-cameras-components-and-lasers/thermal-camera-cores/mwir/

  • https://www.flir.ca/browse/oem-cameras-components-and-lasers/thermal-camera-cores/

Notes

• Kevin Walker

  • project lead, putting hardware into aircraft

  • high level guidance on needs

• Dan

  • sw for AI

  • knows aeryon stuff - Dave Kers….

  • talked with Mihir “extensively” prior month

• they wanted to get right into CONOPs stuff and our requirements

• Boson

  • wavelengths dont line up at all

• Firefly camera series

  • have monochrome versions that let

  • https://www.teledynevisionsolutions.com/en-CA/products/firefly-dl/

    • will work on USB, linux drivers exist

  • wrap in ferrite sheets to avoid emi problems

    • copper tape as well

  • have been using color fireflies

  • need to read datasheets for sensitivity near 940nm

  • 1.6MP on this might be a bit low

    • may be stuff for RPi that does this

    • HAILO is a company that has partnerships for RPi, add on chip that gives AI compute

    • see what we can find, email them and see if they can help

• could be ones for the rpi itself

  • emi issues with that are difficult

  • emi interfering with GPS could be a problem

  • we have GPSes on masts with helps

• concerns

  • if the drone is moving then the beacons will blur

  • 3 meter GPS accuracy is difficult requirement

  • RTK gets good relative

    • it doesn’t give better actual error than normal GPS

  • Surveyed markers might be nice?

    • ask comp organizers about

    • controlled ground reference points is very useful

• global shutter

  • simplifies the math if u have it

  • if u dont have it you need to know:

    • time frame started

    • exposure time

    • these are both called “line time”

• FOV

  • raspberry pi locks it in

  • proccessing limits on the RPi may be a limit

  • cameras for phones always have lenses

• processor

  • rpi uses arm based

  • some dev boards have image processing chips

• frame rate

  • won’t really matter

  • need to process the frames which will take most of the time

• understanding conops

  • first day is map

    • autonomy is big for this day

    • total area is 1km by 1km

    • all beacons will be within 100m circle

    • we can fly for over 30 mins on our batteries

  • second day is use the map to get water

    • this task is separate and not really applicable to FLIR

  • can we do task 1 and task 2 at the same time

    • competition is not setup to allow this