Connexion
Nandu Vadakkath was inspired by a line-following robot built (literally) entirely from salvage materials that could wait patiently and purchase beer for its maker in Tamil Nadu, India. So he set about making his own, but with the goal of making it capable of slightly more sophisticated tasks.
Nandu had ambitious plans for his robot: navigation, speech and listening, recognition, and much more were on the list of things he wanted it to do. And in order to make it do everything he wanted, he incorporated a lot of software, including:
So how did Nandu go about trying to make the robot do some of the things on his wishlist?
The engine uses spaCy to analyse sentences, classify all the elements it identifies, and store all this information in a MySQL database. When the robot encounters a sentence with a series of possible corresponding actions, it weighs them to see what the most likely context is, based on sentences it has previously encountered.
The robot has been trained to follow Nandu around but it can get to know other people too. When it meets a new person, it takes a series of photos and processes them in the background, so it learns to remember them.
Nandu didn’t like the thought of a basic robotic voice, so he searched high and low until he came across the MBROLA UK English voice. Have a listen in the videos above!
The robot has an excellent group photo function: it looks for a person, calculates the distance between the top of their head and the top of the frame, then tilts the camera until this distance is about 60 pixels. This is a lot more effort than some human photographers put into getting all of everyone’s heads into the frame.
Nandu has created a YouTube channel for his robot companion, so be sure to keep up with its progress!
The post Nandu’s lockdown Raspberry Pi robot project appeared first on Raspberry Pi.
We found this project at TeCoEd and we loved the combination of an OLED display housed inside a retro Argus slide viewer. It uses a Raspberry Pi 3 with Python and OpenCV to pull out single frames from a video and write them to the display in real time.
TeCoEd names this creation the Raspberry Pi Retro Player, or RPRP, or – rather neatly – RP squared. The Argus viewer, he tells us, was a charity-shop find that cost just 50p. It sat collecting dust for a few years until he came across an OLED setup guide on hackster.io, which inspired the birth of the RPRP.
At the heart of the project is a Raspberry Pi 3 which is running a Python program that uses the OpenCV computer vision library. The code takes a video clip and breaks it down into individual frames. Then it resizes each frame and converts it to black and white, before writing it to the OLED display. The viewer sees the video play in pleasingly retro monochrome on the slide viewer.
TeCoEd ran into some frustrating problems with the OLED display, which, he discovered, uses the SH1106 driver, rather than the standard SH1306 driver that the Adafruit CircuitPython library expects. Many OLED displays use the SH1306 driver, but it turns out that cheaper displays like the one in this project use the SH1106. He has made a video to spare other makers this particular throw-it-all-in-the-bin moment.
If you’d like to try this build for yourself, here’s all the code and setup advice on GitHub.
TeCoEd is, as ever, our favourite kind of maker – the sharing kind! He has collated everything you’ll need to get to grips with OpenCV, connecting the SH1106 OLED screen over I2C, and more. He’s even told us where we can buy the OLED board.
The post Raspberry Pi retro player appeared first on Raspberry Pi.
One of our favourite makers, Pi & Chips (AKA David Pride), wanted to see if they could trigger a DSLR camera to take pictures by using motion detection with OpenCV on Raspberry Pi.
You could certainly do this with a Raspberry Pi High Quality Camera, but David wanted to try with his swanky new Lumix camera. As well as a Raspberry Pi and whichever camera you’re using, you’ll also need a remote control. David sourced a cheap one from Amazon, since he knew full well he was going to be… breaking it a bit.
When it came to the “breaking” part, David explains: “I was hoping to be able to just re-solder some connectors to the button but it was a dual function button depending on depth of press. I therefore got a set of probes out and traced which pins on the chip were responsible for the actual shutter release and then *carefully* managed to add two fine wires.”
Next, David added Dupont cables to the ends of the wires to allow access to the breadboard, holding the cables in place with a blob of hot glue. Then a very simple circuit using an NPN transistor to switch via GPIO gave remote control of the camera from Python.
David then added OpenCV to the mix, using this tutorial on PyImageSearch. He took the basic motion detection script and added a tiny hack to trigger the GPIO when motion was detected.
He needed to add a delay to the start of the script so he could position stuff, or himself, in front of the camera with time to spare. Got to think of those angles.
David concludes: “The camera was set to fully manual and to a really nice fast shutter speed. There is almost no delay at all between motion being detected and the Lumix actually taking pictures, I was really surprised how instantaneous it was.”
Take a look at some more of David’s projects over at Pi & Chips.
The post DSLR motion detection with Raspberry Pi and OpenCV appeared first on Raspberry Pi.
What’s worse than a clock that doesn’t work? One that makes an “unbearably loud screeching noise” every minute of every day is a strong contender.
That was the aural nightmare facing YouTuber Burke McCabe. But rather than just fix the problem, he decided, in true Raspberry Pi community fashion, to go one step further. Because why not?
Burke showing YouTube viewers his invention
On the back of the clock, alongside the built-in mechanism controlling the clock’s arms, Burke added a Raspberry Pi to control a motor, which he hooked up to a webcam. The webcam was programmed using open computer vision library OpenCV to detect whenever a human face comes into view. Why would a clock need to know when someone looks at it? We’ll come to that.
First up, more on how that webcam works. OpenCV detects when a pair of eyes is in view of the webcam for three consecutive frames. You have to be really looking at it, not just passing it – that is, you have to be trying to tell the time. When this happens, the Raspberry Pi rotates the attached motor 180 degrees and back again.
But why? Well:
A clock that falls off the wall when you look at it
hello #invention #robot #raspberrypi
Burke has created a clock which, when you look at it to tell the time, falls off the wall.
We know: you want your own. So do we. Thankfully, Burke responded to calls in the comments on his original video for a more detailed technical walkthrough, and, boy, did he deliver.
How I made A clock that falls off the wall when you look at it
I dunno why I sounded depressed in this video Original Video – https://www.youtube.com/watch?v=R3HUuf6LGQE&t=41s The Code – https://github.com/SmothDragon/Fa…
In his walkthrough video, you get a good look at Burke’s entire setup, including extra batteries to make sure your Raspberry Pi gets enough juice, advice on how to get to grips with the code, and even the slots your different coloured wires need to go in. And so very, very much duct tape. Who’s going to start a GoFundMe to get Burke the glue gun sticks he so desperately needs? And hit subscribe for his YouTube channel while you’re at it!
The post This clock really, really doesn’t want to tell you the time appeared first on Raspberry Pi.
Lateo.net - Flux RSS en pagaille (pour en ajouter : @ moi) 