AI and Deep Learning for computer vision projects has come to the masses. This can be attributed partly to the community projects that help ease the pain for newbies. [Abhishek] contributes one such project called Monk AI which comes with a GUI for transfer learning.
Monk AI is essentially a wrapper for Computer Vision and deep learning experiments. It facilitates users to finetune deep neural networks using transfer learning and is written in Python. Out of the box, it supports Keras and Pytorch and it comes with a few lines of code; you can get started with your very first AI experiment.
[Abhishek] also has an Object Detection wrapper(GitHub) that has some useful examples as well as a Monk GUI(GitHub) tool that looks similar to the tools available in commercial packages for running, training and inference experiments.
The documentation is a work in progress though it seems like an excellent concept to build on. We need more tools like these to help more people getting started with Deep Learning. Hardware such as the Nvidia Jetson Nano and Google Coral are affordable and facilitate the learning and experimentation.
With his FlexLED project, [Carl Bugeja] is trying to perfect a simple and affordable persistence of vision (POV) display capable of generating “holographic” characters in mid-air. Traditionally POV systems spin LEDs rapidly to create the desired illusion, but that means motors, slip rings, and noise. As the name implies, the goal with this project is to do away with all that and replace it with a self-actuating flexible PCB.
The device is able to quickly move the LEDs back and forth quietly and efficiently thanks to a permanent magnet and magnetic coils integrated into the flexible PCB. With no motors or gears, the whole unit is smaller and less complex than other POV displays. As an added bonus, there’s no danger to the operator or the device should a curious user stick their finger into it.
The last time we took a look at this project, [Carl] had entered an earlier single-LED version into the 2019 Hackaday Prize. Competition was tough last year, and unfortunately FlexLED didn’t get selected as a Finalist. But we’re still extremely interested in seeing the project develop, and we imagine so are you.
The recently completed second version of the display features an improved coil design, eight RGB LEDs and a 3D printed base with integrated magnet. With more LEDs onboard, a single display is able to show multiple characters and even rudimentary animations. A large array of these flapping elements promises to be quite a sight.
But before you get too excited, [Carl] does have some bad news. For one, the cost of building them in small quantities is high, which is always tough for a single hacker trying to iterate a design. Worse, some of the LEDs seem to have died on this prototype already. He says it likely has something to do with the stress of flexing back and forth so quickly, which is obviously a bit troubling. He’s looking to get some feedback from the community, and is hoping to address these issues in the next version.
An ideal application for mesh networking is off-grid communication; when there’s no cellular reception and WiFi won’t reach, wide-area technologies like LoRa can be used to create ad hoc wireless networks. Whether you’re enjoying the outdoors with friends or conducting a rescue operation, a cheap and small gadget that will allow you to create such a network and communicate over it would be a very welcome addition to your pack.
That’s exactly the goal of the Meshtastic project, which aims to take off-the-shelf ESP32 LoRa development boards and turn them into affordable mesh network communicators. All you need to do is buy one of the supported boards, install the firmware, and starting meshing. An Android application that will allow you to use the mesh network to send basic text messages is now available as an alpha release, and eventually you’ll be able to run Signal over the LoRa link.
Navigating to another node in the network.
Developer [Kevin Hester] tells us that these are still the very early days, and there’s plenty of work yet to be done. In fact, he’s actively looking to bring a few like-minded individuals onto the project. So if you have experience with the ESP32 or mobile application development, and conducting private communications over long-range wireless networks sounds like your kind of party, this might be your lucky day.
From a user’s perspective, this project is extremely approachable. You don’t need to put any custom hardware together, outside of perhaps 3D printing a case for your particular board. The first time around you’ll need to flash the firmware with esptool.py, but after that, [Kevin] says future updates can be handled by the smartphone application.
Incidentally, the primary difference between the two boards is that the larger and more expensive one includes GPS. The mesh networking side of things will work with either board, but if everyone in your group has the GPS-equipped version, each user will be able to see the position of everyone else in the network.
I’m writing from a cozy farmhouse just outside of Oxford, UK where we are slowly emerging from a particularly intense Atlantic storm. Some areas have widespread flooding, while fallen tree branches and damaged roofs are countrywide. Our neighbours in the Irish Republic are first in the path of these storms, and receive an especially strong pasting.
In the news following the storm is a merchant ship that was washed up by this storm on the coast of County Cork. The MV Alta is a nearly 2300t and 77m (just over 253 ft) freighter that had been abandoned in 2018 south of Bermuda after a mechanical failure had rendered it incapable of navigation. Its crew had been rescued by the US Coast Guard, and since then — apart from a brief sighting in mid-Atlantic by a Royal Navy polar research vessel — it had passed unseen as a drifting ghost ship before appearing on the Irish coast.
In a very literal sense it had dropped off the radar, but the question for us is how? With the huge array of technological advances in both navigation aids and global sensing available at the end of the 21st century’s second decade, should that even be possible? It’s worth taking a while as land-lubbers to look at how ships are tracked, to try to make sense of the seeming invisibility of something that is after all pretty large and difficult to hide.
His exceptionally comprehensive write-up takes us through the entire process, from creating a rather useful set of 3D-printed brackets for a Pi and camera through deciding the combination of artificial intelligence software components required, to making the eventual decision to offload part of the processing to a cloud service through a 4G mobile phone link. In this he used Cortex, a system designed for easy deployment of machine learning models, which he is very impressed with.
The result is a camera in his car that identifies and reads the plates on the vehicles around it. Which in a way has something of the Big Brother about it, but in another way points to a future in which ever more accessible AI applications self-contained without a cloud service become possible that aren’t quite so sinister. It’s an inevitable progression whose privacy questions may go beyond a Hackaday piece, but it’s also a fascinating area of our remit that should be available at our level.
In the world of additive manufacturing, there’s always need materials being added to the list of potential filaments to use for printing objects. A method of rapid liquid printing of concrete designed by [Anatoly Berezkin] of Stoneflower 3D makes it possible to print a large variety of shapes from concrete while avoiding the negative effects of fast dehydration. The technique is based on an approach to printing polyurethanes, developed by MIT in 2017. This technique requires physically drawing a 3D object within a gel suspension using a chemical curing process. The gel allows gravity to not affect the printing process, as well as helping out with the curinng. Berezkin, an engineer and hobbyist working out of his garage, has published other work including print heads, ceramic printing, and micro printing sets.
One might be skeptical of whether the weight of the material could cause potential collapse during the printing process, or whether it is simply unrealistic to print objects given the time needed for the concrete to dry. Their demo shows the process being done in household items – bowls and tupperware – combining affordable items such as clay, concrete, and sand for the matrix and mortar. The viscous clay is strong enough to act as a good scaffold for keeping the concrete structure in place as it is being printed. As their video demonstrates, at least for simply objects, the process seems relatively fast.
RLPC doesn’t require toxic chemicals or proprietary components such as gels and suspensions. Its immersion of the final printed object in a humid environment is also superior to the standard process of liquid deposition for hardening concrete. Moreover, the process simply requires clay or retarded mortar for the matrix and mortar paste for turning into concrete. It’s advertised as eco-friendly, but just the simplicity of the materials needed for the matrix and mortar make this a promising technique.
In the ebike world, there are two paths. The first is a homemade kit bike with motors and controllers from China. The second is a prebuilt bike from a manufacturer like Giant, with motors and controllers from China, which will be half as fast and cost three times as much. The choice is obvious, and there are other benefits to taking the first path as well, such as using this equipment which now has an open source firmware option.
The Tong Sheng TSDZ2 drive is popular in the ebike world because it’s an affordable kit motor which has a pedal-assist mode using torque sensors, resulting in a more polished experience. In contrast, other popular kit motors tend to rely on less expensive cadence sensors which are not as smooth or intuitive. This new open source firmware for the TSDZ2 further improves on the ride by improving the motor responsiveness, improving battery efficiency, and opening up the ability to use any of a number of color displays. (More information is available on a separate Wiki.)
If you have a TSDZ2-based ebike it might be time to break out the laptop and get to work installing this firmware. If you’re behind the times and still haven’t figured out that ebikes are one of the best ways to travel, here is the proof you need.
Thanks to [coaxial] for the tip! Photo via Reddit user [PippyLongSausage].
Monk AI is essentially a wrapper for Computer Vision and deep learning experiments. It facilitates users to finetune deep neural networks using transfer learning and is written in Python. Out of the box, it supports Keras and Pytorch and it comes with a few lines of code; you can get started with your very first AI experiment.
