AI Contribution towards Clean Water

By Smita Kumari |Email | Jun 7, 2018 | 11973 Views

Today we face most of the disease because on lack of clean or purified water. With the growing population and increasing pollution, heath related problems became major concern. The development in AI now capable to detect dangerous bacteria in water. 

Community face biggest challenge to test water as it requires expensive equipment and lot of manual analysis that is difficult for every citizens and ensure that they are drinking safe and purified water. So, researchers have identified the potential benefit that AI can offer in checking the quality of pure water.

Peter Ma, an intel software innovator in Intel AI Academy is working on Clean water AI project and has developed an effective AI-system that can test the purity of water and detect dangerous bacteria. It is very good and effective test system for those who face continuous problem of getting clean water.  This AI-enabled system uses pattern recognition and machine learning to identify dangerous bacteria in water sample and prevents disease, this in turn results in improved lives.

In clean water AI project, Peter took advantage of Intel AI DevCloud to teach AI model to perform testing of water purity in real time. Intel provides both software and hardware needs in AI to perform its function.

AI Test System to check the quality of Pure Water
The AI system to test the purity of clean water is an inexpensive method. It comprises of four key parts:

  • Intel AI DevCloud
It is a free cloud compute available for the members of AI academy which Peter has used to train their AI model. It provides access to cloud-based intel Xeon scalable processor for machine learning and deep learning training.
  • Digital microscope
It is used to view the water sample dropped on a slide to detect bacteria.
  • Computer
A computer running Ubuntu operating system used to do a real time analysis of water ample with the help of trained models.
  • Intel Movidius Neural Compute Stick
It is an USB based device that is used to accelerate the inferences of the tests.

Lets have a look how it works:

This AI based water testing solution uses Convolutional Neural Network to analyze the water sample and identify the shape, size, color and density of bacteria. The solutions also include a website that locate clean and contaminated water locations to the user.

At the current point of time this system can detect only Escherichia coli and bacterium that causes cholera but it has potential to extend its capability in identifying bacteria of different physical characteristics. Peter Ma is still working to trained his AI model and extends its capability. 

Source: HOB