(Photo : Image from Kickstart) Depth Sensing AI Camera OAK-D Lite Gets Smaller and Smaller | Full-Color 4K, Greyscale, Onboard AI Machine Vision Processing
The new OAK-D is reportedly an open-source, full-color depth sensing camera that is embedded with AI capabilities. Currently, there is now a crowdfunding campaign going on for a brand new lighter version that is now smaller called the OAK-D Lite.
Oak-D Lite Specs
As seen on Kickstarter, the new model actually does everything that the previous model could do as well. This is by combining machine vision along with depth sensing and even an ability to run some very complex image processing tasks all on-board. This will then free the host from any of the usual overhead that is needed.
According to Hackaday, the said OAK-D Lite camera is now actually a number of elements all packed together in a single package. This includes a full-color 4K camera, two different greyscale cameras for stereo depth sensing, and even an onboard AI machine vision processing that works with Intel’s Movidius Myriad X processor.
Open-Source Platform Depth AI
Tying everything together is a new open-source platform that is known as Depth AI that will reportedly wrap the camera’s functions and capabilities together into one unified whole. The main goal is to be able to give embedded systems access to certain human-like visual perception in real time.
At its core, this means that it would include detecting things as well as identifying where they are in physical space. It reportedly does this through the combination of the traditional machine vision functions just like the edge detection and perspective correction along with depth sensing as well as the ability to plug in pre-trained CNN or convolutional neural network models to work on complex tasks just like object classification, pose estimation, or even hand tracking in real-time.
How Does the OAK-D Lite Work?
The device is reportedly used like a USB, practically. This means that users will need to plug the OAK-D Lite into a host that is running any OS and the team has actually been putting in a lot of work in order to make this as easy as it can be. As of the moment, Google AI department is getting a class-action lawsuit for a whopping 1.6 million confidential medical records of NHS patients being used.
With the help of a certain downloadable application, the hardware can then be up and running with certain examples in just about half a minute. When integrating the device into some other projects or products, users can do this through the use of Python with the help of DepthAI SDK.
DepthAI SDK provides functionality all with minimal coding as well as configuration which is reportedly more advanced users. There is, however, a full API for low-level access as well. This is since the vision processing is all being done on-board that even a Raspberry Pi Zero can be used in order to effectively be the host. WhatsApp is reportedly now encrypting its backups for stored data on Google Drive and iCloud which can be accessed with a random 64-digit key.
Written by Urian B.