Is it Doable to use IMAGE PROCESSING for waste segregation?.

I am a senior Computer Engineering student from the Philippines.
I wanted to know if ARDUINO is Capable for image processing

1 Like

No, Arduino had no processing power for image processing on both resource and speed.

Well first of all which arduino are we talking about?
If you mean the ones that use an avr microcontroller running at 16mhz then the answer is it can do some basic stuff.

thank you sir.

what micro controller should I use sir? because we are not allowed to use a PC on this project.

will raspberry pi will do for this image processing?

Without any components to support it you have a maximum of 2K of memory. Which you would know if you had done any research.

So providing your image is say 1K then there are things you can do.
The very first TV (John Logie Baird) used the same sort of resolution 32 by 32 pixels.

will raspberry pi will do for this image processing?

Yes you can do real image processing on a Pi.

Yes you can do real image processing on a Pi.

thank you sir..

In my project, the use of image processing is to segregate trash for "biodegradable" and "non-biodegradable"
I think that it is doable to segregate trash through image processing. but need further research and consultations.

I think that it is doable to segregate trash through image processing

Ah the optimism of youth.

Good luck.

It sounds like you have not purchased anything yet. What you want to do sounds like it will use advanced cpu intensive algorithms. Is that correct? If so get a beaglebone black or raspberry pi. If you want to use an arduino get an Intel Galileo or maybe Arduino Yún. I am not sure about the arduino offerings but they may work.

Ah the optimism of youth. Good luck.

TRANSLATION:(?) The image processing required for your application probably exceeds the hardware capabilities of the equipment you are planning to use. The software is another issue.

Alternative translation:

Image recognition capable of identifying and recognising a sequence of arbitrary objects comparable to trash flowing past a camera will be extremely difficult, even assuming you chose a sensible platform that had ample resources. If you have a team of postgrads with image processing experience and several years to spend on it, you might have a chance of getting somewhere. Given the requirement that you do not use a PC, I suspect you will find that this is substantially harder than the sort of project your tutor envisaged you tackling.

@centre_fields, do not cross-post. Threads merged.

Grumpy_Mike:
Ah the optimism of youth.
Good luck.

Real translation.

I spent many years doing image processing research in the 80s. I designed and built a real time frame grabber for a TRS80 computer for $100 when the only such systems that existed cost hundreds of thousands of dollars. Amongst other things I used it to evaluate the efficacious of brain surgery in correcting facial tics. I have had students working for me on image processing projects. One of which was to look at an analogue clock face and use image processing to tell the time. Sounds a achievable aim? The student failed to get anywhere with the project.

My personal opinion is that you don't stand a hope in hell of achieving anything useful. But if you never try for your self and just rely on the past experience of old farts like me no one will make any progress. You will only realise how difficult the task is once you try and do it, so go ahead and try. You might find you can get something to kind of work, and the experience of trying will be invaluable.

Good luck? - Yes good luck.

I agree. It's hard. But reading a clock ? That's not very hard.

What we are planning to do is to segregate Papers, Plastic Bottles, Tin Cans, Plastic.

not as broad as what I wanted to do.

It is only limited image processing. I know that it is not possible to process all type of trash by image processing.

Get a raspberry pi or a beaglebone black. Both are capable boards each with pros and cons. The reason I suggest them is due to high performance,existing camera solutions, availability and low cost.

michinyon:
I agree. It's hard. But reading a clock ? That's not very hard.

Have you ever tried doing it? It is a lot harder than you might think on just first hearing of it.

Mr_arduino:
Get a raspberry pi or a beaglebone black. Both are capable boards each with pros and cons. The reason I suggest them is due to high performance,existing camera solutions, availability and low cost.

I would endorse that especially with regards to the Raspberry Pi. It will take a web cam or it's own Hi-res camera.

Im still skeptical.
but I am partially sure that I am going to use raspberry pi for image processing.

I wanted to ask again if it is doable to use IMAGE PROCESSING for simple waste segregation?

If you have a good algorithm to do such. What kind of waste are you trying to segregate? What makes it different visually?

http://blog.oscarliang.net/raspberry-pi-face-recognition-opencv/

Mr_arduino:
If you have a good algorithm to do such. What kind of waste are you trying to segregate? What makes it different visually?

Im going to segregate "Plastic Bottle, Tin Cans, Papers. just common recyclable stuffs to biodegradable waste.

e.g : Recyclable trash and Biodegradable trash.