Leverage technology to support sustainable environment goals. For example, counting the number of salmon during a salmon run is very important, and a sensor node with computer vision capabilities at the "Edge" powered by technologies such as the Intel Neural Compute Stick and open source software provides a good option. In addition, the distributed cloud integration with DeepcloudAI (https://deepcloudai.com/) for the back-end provides low latency and better reliability.
What it does
Integrates leading technology components to create a powerful combination to address environmental challenges. Specifically, uses a camera board on a Raspberry Pi for video capture, processes that video through the NCS, posts the analysis output to the asp.net core based website hosted in a docker container in the Deepcloudai platform.
How we built it
Relentless hacking to get things to work, and to get them to work together. Downloaded and installed the NCS DK needed on an Ubuntu laptop and a Raspberry Pi. Used the asp.net core sample to start the app starting with this URL https://github.com/dotnet/dotnet-docker/tree/master/samples/aspnetapp Modified the home page to refer to Deepcloudai The docker image has been uploaded to: https://cloud.docker.com/repository/docker/sudhirkshirsagar/aspnetapp The image was then deployed into the Deepcloudai platform using the free credits provided.
Challenges we ran into
Incompatibility of software components
Accomplishments that we're proud of
Got an Intel Neural Compute Stick (NCS) to work with our Ubuntu desktop and a Raspberry Pi. Created and deployed an asp.net core app in a docker container into the decentralized Deepcloudai platform.
What we learned
A quick start into leveraging the NCS for edge-based AI, and using a decentralized cloud through docker container deployment. Also learnt more about the use of blockchain in provisioning services.
What's next for Environmental Sensing: AI, IoT and Decentralized Cloud
Explore how "Blockchain" can be used in the managing an environmental sensing network with a distributed cloud. There are examples of how Solution architectures like Supply Chain Track and Trace can be implemented with blockchain in a central cloud. https://azure.microsoft.com/en-us/solutions/blockchain/ https://github.com/Azure-Samples/blockchain/blob/master/blockchain-workbench/iot-integration-samples/ConfigureIoTDemo.md We would like to explore those concepts in a decentralized cloud like Deepcloudai.