Arduino Prototyping Inputs #63: Image Tracking with the PIXY! Camera
HTML-код
- Опубликовано: 3 сен 2019
- Image recognition and movement is made simple by using the PIXY! Camera system and Arduino. This camera allows for advanced color tracking, pan & tilt, and a variety of other powerful features.👉 Leave a tip via cryptocurrency if I've helped you out:
Bitcoin: bc1q2whh7h7nddf6lzzsns6phxdf2pyr6z0pvgy98t
Doge: DDgFpxEsBTdjgJDg3vxwuagia9UU5UmSVL
Etherium: 0x952e75561491951C50561b76eCFFFE1AC493d495
XRP: rPVaL4TNr3aMPhwgYsWZdgirM99S9QnfCA
Polkadot: 15pNwwPAsMwi59cMJC441P6KwYzCsq9WqQu156FF8KffRaee
PIXY! Camera -- amzn.to/32Y0yyC
NEW! Learn to Solder Course (Free to LinkedIn Pro users) - bit.ly/2ILSzx3
Prototyping Supplies:
• Inexpensive Arduino Compatible Kit -- amzn.to/2Dfqs5U
• A More Deluxe Starter Kit -- amzn.to/3g9nFtR
• Jumper Wire Kit - amzn.to/3f9rtK4
• Breadboard Kit - amzn.to/3hIEPib
• WaveShield Kit -- amzn.to/3g9vMq6
• Motor Shield Kit -- amzn.to/3hISQfL
• Adafruit Circuit Playground Kit -- amzn.to/39yBfo9
My Camera Equipment:
• Primary Camera -- amzn.to/3fewxgA
• Camera Kit -- amzn.to/3jQvaI8
• Primary Lens -- amzn.to/3jQvaI8
• Macro Probe Lens -- amzn.to/3jQvaI8
• Main Light -- amzn.to/2X1SwB6
• Secondary (RGB) Lights -- amzn.to/3jQvaI8
• Mini Lights -- amzn.to/3jQvaI8
A few other FREE sensor tutorials you might like: (I have 100's!)
• Cellular Data shield - • Arduino Prototyping Te... z
• Capacitive Sensing - • Arduino Prototyping In...
• Luminosity Sensors - • Arduino Prototyping In...
• Thermocouples - • Arduino Prototyping In...
• Sensing Vibration - • Arduino Prototyping In...
A few other sensor tutorials you might like: (I have 100's!)
• Cellular Data shield - • Arduino Prototyping Te... z
• Capacitive Sensing - • Arduino Prototyping In...
• Luminosity Sensors - • Arduino Prototyping In...
• Thermocouples - • Arduino Prototyping In...
• Sensing Vibration - • Arduino Prototyping In...
And some playlists on prototyping with Arduino:
• The Basics - • Arduino Prototyping: T...
• Inputs - • Arduino Prototyping: I...
• Outputs - Arduino Prototyping, Techniques - • Arduino Prototyping Ba...
• Complete Prototyping Series! - • Arduino Prototyping: T...
Instagram - / marcdevinck
Twitter - / devinck
Don't forget to like and subscribe if you'd like me to make more videos!
My company, Panoptic Labs - www.panopticlabs.com 
Хобби
Awesome overview video
your channel is a hidden gem!
Thanks!
Wow, I didn't know it is that simple. I mean, it is not that difficult as I thought it is. Thanks for explaining how it works.
The PIXY! Cam is a fun addition to your prototyping lab. It works really well too!
Can I attach two step motors to control the movement of the camera (up down and side to side) to follow an object ? I would have to create some physical base for the camera to sit on obviously. Let me know if you have any ideas!
Yes, you can use 2 servo motors (not stepper motors, they require a lot more circuitry). You want a pan & tilt base, which are available and easy to implement.
Informative Video!! I was looking for a camera for a project when I stumbled on this video. I want to capture the positon of a laser pointer on a piece of white paper. if the Pixy2 reads 60 times per second that would be 16.6 ms per read. I was going to have a different Arduino turn the laser pointer on for 100 ms ( adjustable ) So with the laser pointer on time at 100ms it should read it 6 times, I was going to average those and come up with the x and y location. Before buying a Pixy2 I thought I would ask, do you feel that it would be able to do this?
Thanks for any info
Yes, that sounds OK, but the frame rate can vary (from my experience) so if you are tied to those #'s, you might want to look at something more robust. I still think your example would work. The PIXY! is very capable.
what is the limitation of this sensor. it seem working well. is it the quality of the camera? is it able read the infrared in warehouse or watching temperature of machine. if it find something then sending alarm.
it is just project for factories, heavy machines that working 24/7.
The PIXY Camera system is most likely not the best for that specific use. It's a great camera, but I think you would be looking for an IR thermal camera, which is readily available on the market as an accessory to the Arduino.
Can it recognizes more than one object like orange, water lemon and banana?
The new Pixy2 can track several (lots!) objects at once, do line detection, road sign detection, and even know when it comes to intersections of lines. It's an awesome little piece of hardware. So to answer your question, yes, the Pixy2 can track multiple objects.
Question: If you disconnect the internet from it, will it still work on tracking moving objects without its connection to the internet, or should it be connected to the internet in order to continue tracking the movement of objects, and if it works, can we make a project to track an object using Arduino Uno
The PIXY! Cam, along with the Arduino in not connected to the internet, and works standalone.
@@MarcdeVinck It’s HD or Full HD.
Models always are updated and specifications change, especially in DIY electronics. Always refer back to the manufacturer's website for the current information and models. pixycam.com/
great video! how many targets can it "remember"?
They have anew model! Please check out their website for the specifications -- pixycam.com/
@@MarcdeVinck Is there a way to purchase the v1? I saw the v2 but it's a bit out of budget
@@l8dawn There may be some online retailers with older stock. Might also want to reach out to them.
I just curious, where did you get the zip file? I cant locate in the description box
It's on their website -- pixycam.com/downloads-pixy2/
I'm looking for a way to have arduino take a screenshot and identify a matching screenshot taken previously and then run a script based on which screenshot has been matched.
Would pixy cam be capable of this?
Is there a better solution?
You'd be much better off using a Raspberry Pi running TensorFlow. You can do image comparisons, recognition, etc. Good luck!
@Marc de Vinck thanks for the reply Marc, I'll check out the tensor flow. My proposed application is to find a way to automatically identify which video game is selected in a multi game box so a joystick can shift into either 4-way or 8-way as the selected game requires.
Mame setups running on personal computer, raspberry pi, etc, can be programmed directly to do this but boards/boxes are not programmable in same way. My idea was to split the video output going to the monitor plugging one into a board that could identify based on matching screenshot at time game is launched to recorded screenshots in memory.
@@trailb4u If you are already on a computer (or Pi), why use image recognition? The system knows what game was launched, just switch the controls as needed. It should be fairly easy to implement a lookup table of games and control methods.
@@MarcdeVinck Hi Marc, I'm looking for a solution for the boards that are not a computer or pi, (a 60 in 1 arcade game board for example). In a designated area of the screen, the actual name of the game appears in large print before pressing player 1 or player 2. So a text reading program would do the job in this case I think. The hardware cost would probably need to be in the 25 dollar range for this proposed workaround to be viable. I'm just checking to see if there is an off the shelf, cost effective solution.
@@trailb4u Oh, OK, that makes a lot more sense. TensorFlow on a Pi would be a good choice IMHO.
How it can track and lock automatically any object using camera and aurdino?
No 100% sure what you are asking, but that is a much larger project than you would most likely use an Arduino and AIXY! Camera. I suggest you look into Tensor Flow and a Raspberry Pi to get started with object detection and tracking.
Yes im doin project and facing problem that how to lock moving object ..if u can help plz do me favor
@@shaikhjunaid5577 I would look into TensorFlow running on a Raspberry Pi. Good Luck! Sounds like a fun project!
can it scan and recognize the objects on mobile screen?
Objects as in a shape and/or color, yes. But object recognition like a face, dog, or chair, you'd be better off running TensorFlow Lite on a Raspberry Pi. Needs more power!
@@MarcdeVinck okay thank u
if i want to make real product. Could you please suggest better sensor than PIXY?
You can use the PIXY! Camera for a "Real product"! It's great! I can't suggest a better sensor without knowing all the details of what you define as better? Larger image, faster shutter, lens, etc. There are lots of cameras available on the market for DIY electronics. Head over to your favorite online retailer and take a look at the specs to find one that matches your requirements. Just keep in mind the real value of the PIXY! is not the camera but the additional hardware and software that supports the camera.
Can it recognise angle between two objects or straight line
It could, but not on it's own. For something like that, you'd be better off running Tensor Flow and a camera.
@@MarcdeVinck oh i wanted something that could recognize and object's x and y co ordinates and angle between them with respect to hirizontal and vertical line
@@KKMaity It's possible, but there are a few things to consider that may need more power and/or integrate some AI. What do yo what to recognize? Is it a color different than the background? That's not too bad. Or is it a specific shape or "thing" that would need something like TensorFlow to run an instance model to compare and identify a video feed to know objects. Also, resolution and size are major concerns. How accurate? Of it is recognizing 2 yellow objects on a clean background (no other yellow) and you want to calculate the angle between +/- 5º, I would think you can do that.
@@MarcdeVinck suppose there is a background of darker colour and on top of it two circular objects..both of the objects have different colours and also lighter than background (example background black two objects red and yellow) I want to know x and y value of the center of those circles.. and want a robo arm to reach at that position..
@@KKMaity Seems possible. Again, depends on the resolution required, accuracy, and speed required. Also note that there is no depth, so again a depth camera running TF might be best, but you may be able to get some sort of depth with calculating the size of the objects?
Your videos sound volume is low, please raiia bit
😅Hey i want a help on this type of project how can I contact you iam a engineering student from India can you help me out 👀
Sorry, I don’t do any consulting. Check online for some local groups. They may be able to help.
@@MarcdeVinck I need professional help it's last stage of very innovative project
@@akash_here If you need professional help, I suggest you look for a local Industrial Design company or Product Design company to help.