Team Name: OneManArmy
Short Story:
After a long break from my last project which been years ago I've decided to visit hacksters.io to view my profile and what's new, I came up to the competition page by accident and saw the open.cv competition which was a dream to use in the past for a personals robotics project but I had no time to start such project in the past.
This competition gave me the motivation I needed and to challenge myself and start again the project days, which feels great, keeps my mind focused and my imagination active.
Project:
This project came in my mind Multifunctional Visual Device (MVD) a combination of my past dream project with open.cv so I've decided to give a try and challenge myself in this competition.
The purpose of this device is to be used in multiple fields:
- Agriculture to assist and robotize motor/electrical vehicles;
- Industries as well to robotize equipment;
- Daily driving as road assistant or future car automation;
- Architecture to review the old or new building,
- Security systems to monitor specific flows, detect or allow access;
- and in Search and Rescue as Caves, Ground, Mountains or Maritime
As this is early stage to insert more information about this multiple purpose device (MVD) it's a work in progress so stay tuned for more updates.
Toolsand Waysto build it:
Homemade PCB, Electronics, Arduino, Python, Open.cv, Dremels & other general tools, 3D printers (probably have to buy one).
19/08/2023 Update,
Exiting day today, I received few of my project components, the 4 FHD 130 Degree cameras which I'll be using
A quick test and they are working perfectly (only 3 as I don't have another usb port) now I need to design my own PCB and camera attachments to have a full 360 view. Time to work hard.
Ow almost forgot, I bought a 3D printer as well to design my own prototype casing and final assembly of the device
This will be a future update.
Realising today that I'll not be able to use multiple cameras via a single USB as the bandwidth of the cameras exides the usb bus capabilities
But that doesn't stop me, an idea is to decrease the bit rate fps and use gray scale as it will use less Mbps and I might be able to use at least 2 cameras per hub or plan B to use a SATA hub with a custom 20 pin port will see after testing phase.
Meanwhile will design a temporary 3D printed frame to aligne the cameras and see how the future device will look like.
Small update of the day
Successful alignment of theoretically 4 cameras flip and rotate, after the prototype case for the 4 cameras together to create a 360 view, ofc some cropping and adjustments will be required and some lenses.
Update 22/08/23Problems with USB supporting FHD or HD per single driver with multiple cameras so problem in progress.
Meanwhile 1st prototype to print camera supports, had other ideas but design changes.
This is for testing purposes to make my life easier. Printing results later in the day.
Later that day...
Updates23/08/23
SSome design updates, decided to make a final design shape which will be a sphere with a cool exterior texture (surprise design), added some camera supports which will melt them and fix the cameras on the mounts, as well added holes to pass future cable management through the mounts.
Those tubes are lasers, each camera will have its own laser to measure the distance of the objects and theoretically use the distance data to create a 3D output, as well the lasers will be spinning 360 degrees internally but this is under progress as I need to see how to do that using some motors and stepper sensors.
Update 26/08/23Could not wait to share my happiness, I just received one of the important components, lasers(picture bellow).
Time to design some supports, attachment and gears to spin the lasers, but I'll need some very small motors well this will be an update for later. (Time to work)
Update 28/08/23Been working on a prototype laser support attachment to the rest of the structure, enjoy the pictures :)
This is not the final design as I want to add some rotational sensor, 3D print some gears and add small motors.
Note: I added 20 to 30 degree angle to the sides to not bother the camera lense and output of the video.
Short recup: Cameras tested, lasers working, the scheleton structure is almost ideal to have a base of the actual entire device of course still a lot to design till the finished desired outcome.
Next step I believe is to finally decide what kind of single port interface I will have and probably run some tests with Arduino to attach the small sensor to determine the angle of the lasers and ofc design some gears to attach them.
I know you might say "but what about open CV development" well it will come soon enough, the most important right now is the actual device itself to have it developed based on the open CV desired outcome, it's easier to work on code when you have a stable base structure. (More updates in the future)
Update 29/08/2301/09/23 Been working on alignment of components inside the sphereBuilding gears for the lasers, I gave it a try and 1st try was a failure, made gears smallen than what my printer can handle, so I designed bugger gears.
Wanted to use small motors but because of some investigation about how they going to work decided to use other motors a bit bigger but precision matters, so I changed the motors, it will be a battle for remaining space but there is still parts to improve.
Another I start working on is a platform for power distribution PCB for the lasers and motors, as well the platform will theoretically hold the sensors which will determine the rotation of the motors in degrees.
A work in progress is as well clamp holders for the lasers, support platform/socket for the motors and motor gear distribution and shaft to connect the rotational sensor reader.
More updates in the future.
Update 02/09/23I think that's the most difficult and time consuming component of this 3D design project, took me ages to build.
About 5 hours of measurement, precision alignment up to 0.5mm precision so the components would fit.
Well I present you the screw gear with bottom motor mount, a top sensor attachment shaft and theoretically should spin 2 other gears. (Let's hope it work from 1st try because I'm done with it :) )
And this is how it should look.
Time to 3D print it and see what adjustments are required..
to be continued:)
Comments