Starter Project: Camera and Live Feed

[im_a_kobe]

Hello folks,

First of all, a disclaimer: I am an absolute beginner and I'm seeking a way to actually begin. If this is the wrong page/protocol for this, please kindly direct me to the right place. Else, read on!

I would like to create a very simple concept, but likely a really complex execution: a camera sensor and a screen that displays what that camera shows. I hope to create it from scratch, so not using Arduino/RPi modules. This is just for fun and learning so there are no time constraints and I'm not afraid of programming or PCB design. Although I have no electronic experience besides a couple of physics classes in college, I would love to take on this challenge.

What I do know is that i can order some components separately, ie. an oled screen with a parallel interface and a CMOS sensor of some sort. My very limited logic states that i will probably need a microcontroller of some sort and a way to power the device. Whether or not this information is actually correct I don't actually know, so I'm hoping to get some direction from you smart folks!

Anything you can think to assist me would be helpful, but here are some questions:

-What components should I look at getting? I don't need storage, I just want to display a camera feed in one enclosed system, like a phone does when you open the camera app.
- What should I expect to do/learn to do when putting these components together? Ideally no soldering since i have terrible hands for such things, but I'm willing to learn anyway and soldier through.
- What are some good resources that can help me achieve this particular goal? I don't want to become an electronic engineer, or an expert in PCB design but I'd like to know just enough to create this one thing.

Cheers,

TL;DR: I'm a beginner who learns best when i have a tangible goal, such as creating a device that shows a camera feed. I would like to create this device and need some help starting out.

 

[Bluejets]

Complex place to begin.
Can you turn on an LED within specs?
Most don't even bother as the system you describe is already available at low cost, assembled and working, usually uC controlled for a mass of options.

An example is a Banggood camera for FPV which costs $7.00 and associated display for an extra $20.00.

Add a wireless transmitter and receiver ( 40 channel) another $15.00.

Point being, one can cover a LOT more learning territory by using pre- made modules AND at a fraction of the price with a semi-professional finish and reliability.

Welcome to the 21st century.

 

[im_a_kobe]

Can you turn on an LED within specs?

No, but I'd love to learn!

Most don't even bother as the system you describe is already available at low cost, assembled and working, usually uC controlled for a mass of options.

An example is a Banggood camera for FPV which costs $7.00 and associated display for an extra $20.00.

Add a wireless transmitter and receiver ( 40 channel) another $15.00.

BangGood FPV accessories are an excellent place to begin thank you. I'll definitely explore the options they have on there.

Point being, one can cover a LOT more learning territory by using pre- made modules AND at a fraction of the price with a semi-professional finish and reliability.

I'm not in a rush, nor am I trying to have an amazing product at the end, I would just like to be able to purchase some base components, put them together and have enough low-level control to adjust things like refresh rate, latency etc. That's why I'm avoiding premade systems, since I figure that sourcing screens and camera sensors themselves and delving into PCB design will allow me this....I could be wrong ¯\_(ツ)_/¯ I understand that it's difficult, but if it wasn't it wouldn't be challenging.

I relish the challenge.

Thanks for your help!

 

[kellys_eye]

I am an absolute beginner

but likely a really complex execution:

I hope to create it from scratch,

Three statements that really don't mix! That's a bit like "I know how to use a knife and fork but I really want to do open heart surgery".

You need to know more about the basics of what you're wanting to achieve - first principles if you will and, certainly, the knowledge of the protocols involved (as there are many).

What I do know is that i can order some components separately, ie. an oled screen with a parallel interface and a CMOS sensor of some sort. My very limited logic states that i will probably need a microcontroller of some sort and a way to power the device.

You could start by chosing your components - download their data sheets and learn what they mean and how they do it. If you had a basic diagram illustrating the signal path (left-to-right, lens, camera sensor, interface, display - then there will be the power supply(ies), the controller(s) etc) you can go off and find the matching bits.

The datasheet for the camera sensor will, for example, tell you about the required interface, the protocol of the signal, the lens requirement, the supply requirement etc etc and take it from there.

But even finding the right camera sensor needs more thought.... resolution? mono or colour? sensitivity? serial or parallel (or some other interface)......

What you've chosen to do will simply reveal that you can't avoid learning the basic principles first.

 

[im_a_kobe]

Cheers @kellys_eye

Three statements that really don't mix! That's a bit like "I know how to use a knife and fork but I really want to do open heart surgery".

Haha aye it's a bit like that, but I'm looking for guidance from the surgeons to tell me that i can't use a butter knife, I should use a scalpel, that the first incision is almost always vertical and their rule of thumb is to never sneeze. There are no shortcuts in life but the internet must have some.

You could start by chosing your components - download their data sheets and learn what they mean and how they do it. If you had a basic diagram illustrating the signal path (left-to-right, lens, camera sensor, interface, display - then there will be the power supply(ies), the controller(s) etc) you can go off and find the matching bits.

The datasheet for the camera sensor will, for example, tell you about the required interface, the protocol of the signal, the lens requirement, the supply requirement etc etc and take it from there.

But even finding the right camera sensor needs more thought.... resolution? mono or colour? sensitivity? serial or parallel (or some other interface)......

This input is much appreciated. I now know to look for a camera sensor and begin deciphering it's datasheet. I would like a color sensor, with a high bandwidth interface (looking to learn about high refresh rates etc. - diving right into the deep end).

What you've chosen to do will simply reveal that you can't avoid learning the basic principles first

I am not afraid of failure, I'd like to use this method to learn those basic principles. For context, I am a mechanical engineer with a decent set of programming, hardware/software development and soldering experience. As such I feel confident in my ability to pick up the basics as i go and retain the principles that i need. I'd just like some direction.

I'm out of my comfort zone in the EE field and would like the opportunity to change that by creating something I'm interested in. If the ultimate answer is that I need to spend some months with an Arduino and then revisit I'm open to it but I'd like to explore my options with this particular dream.

Thanks again! Looking into the sensors, will see if that kicks off the project for me.

 

[Bluejets]

Personally, I'd go straight past electronics to quantum science.

 

[im_a_kobe]

Personally, I'd go straight past electronics to quantum science.

I tried that.. but i was so uncertain about the principles that they kept disappearing before i had a chance to observe them.

 

[kellys_eye]

They'd have changed anyway so why bother?

 

[BobK]

If you wanted to learn mechanical engineering, would you start by designing a car from scratch?

Bob

I hope to create it from scratch, so not using Arduino/RPi modules.

So, are you going to start with sand?

Bob

 

[im_a_kobe]

They'd have changed anyway so why bother?

Just hoping that I'd be looking the other way when 42 particles hit me in my Higgs-Boson and i accidentally solve the worldwide towel shortage, thereby making the galaxy infinitely safer for everyone.

If you wanted to learn mechanical engineering, would you start by designing a car from scratch?

Probably not, but i wouldn't discourage someone willing to take up the incredibly daunting task if they had the time and the means. Sure they are starting backwards, but through the chassis design they'll realize that they probably need to learn quite a bit of solid mechanics, but probably still go pretty far on their own through trial and error. Through engine design they'll realize they probably need to learn quite a bit of thermodynamics if someone is there to help them not blow themself up creating an ICE in the garage, a person does not need to learn the law of entropy to be taught the 4 stroke engine cycle. I don't understand all the snark, this method of learning is very valid if there is a reasonable level of support. Just look at all of the little kids learning rocket science and orbital dynamics by building rockets in KSP. You can understand F=M*a through practical means without needing to sit through a class of all of all Newton's laws.

At least help me to understand why it's difficult. Something like "well you can't use a cmos sensor with an fgpa because the flux capacitor will explode, normally people use the purple stuff but it's super expensive and one needs to calculate the third law of kerbin using the back of their electrician certificate before it works" and i can at least try and decipher that as I need to learn how to use the purple stuff.

So, are you going to start with sand?

Again, not opposed to Arduinos, and I'm welcomong any nuanced reasoning/suggestions on taking that approach. I'm just under the impression that they may be limited in what they can handle for throughput etc. due to them being designed for other things. I would like to follow a path similar to the one outlined in this thread:

https://www.microchip.com/forums/m/tm.aspx?m=402377&p=1

Seems like they are able to get somewhere using PIC32 microcontrollers. Any thoughts on going that direction?