Importance of water for life:
Our earth – the blue planet.With 1.386 million km³ water a unique place in the universe. Our basic need is water to support the survival existence in the environment. Water plays an essential role in nature. Not only as a drinking source but it is also important for whether cycle, agriculture process, in a plant life cycle and dwelling place for watery creatures.
Statistics of water in the present age:
Our global water reserves state that only 75% of the planet earth covers by water, that too in liquid, gaseous and frozen state. But humans can not use these forms of water. As 97% of the planet’s water resources are saltwater. Very few percents of water for about 0.649% is considered as fresh water. Therefore, there’s a need to research and develop innovative solutions for the treatment of 98.5% of water available worldwide but cannot use directly. So, a total of 98.5% of water availability covers water from salt bodies(97.6%), groundwater (0.8%) and surface water(0.1%). Remaining 1.5% is frozen in glaciers, ice and snow. Problems facing due to the impurity of water: Water bodies gets contaminated due to the pouring of industrial wastewater into the rivers, lakes etc. Industrial water consists of several toxic chemicals and bacteria which pollutes the river. Using such water for drinking purpose and for agricultural process led to severe water-borne diseases like cholera, diarrhea.
What is the Challenge:
As the population is increasing day by day, getting clean water is just as impossible. We need to do several tests and analysis on the water to check the quality but requires expensive equipment. So, it’s a challenge that how to detect the contaminated water from the clear water to use for drinking and agriculture purpose with minimal cost that too for daily purpose. And to supply clean water to every human to live a fit and healthy life.
Best way to tackle the challenge: In the present technology era, we can solve every kind of problem by applying Artificial Intelligence techniques.The Idea of Inexpensive AI-Driven test system to detect bacteria using advanced pattern recognition and machine learning in water developed by Peter Ma. He is an Intel Software Innovator. He used a digital microscope and connected to a computer having Ubuntu OS.
According to World Health Organization, 2017 “Every minute a newborn dies from infection caused by lack of safe water and an unclean environment.”
Intel AI Enabling Technologies used:
- Intel® AI DevCloud – Intel provides free cloud computing for their AI academy members. It is powered by Intel® Xeon® Scalable processors, used for machine learning and deep learning training and inference computation.
- Intel® Xeon® Scalable processors – Can optimize performance, support to hybrid cloud infrastructure can do efficiently feasible insights,hardware-based data security.
- Intel® Movidius™ Neural Compute Stick – Deploys Convolution Neural Network (CNN). Supports both the Caffe and TensorFlow frameworks, popular with deep learning developers.
“Intel provides both hardware and software needs in artificial intelligence—from training through deployment. For startups, it is relatively inexpensive to build the prototype. The AI will be trained through Intel® AI DevCloud for free; anyone can sign up. The Intel Movidius Neural Compute Stick costs about USD 79, and it allows the AI to run in real-time.” – Peter Ma, Software Innovator, Intel® AI Academy
How to identify bacteria in the water sample?
Clean Water AI test system composed of simple, inexpensive and offline components: Intel® AI DevCloud used to train AI model using Caffe. A digital microscope is available for USD 100 or less. Used for capturing the image of a water sample dropped on the slide with a pipette. A Computer running with Ubuntu operating system to run real-time analysis and find contaminated water having E.colli bacteria using the captured image of the water sample. An Intel Movidius Neural Compute Stick is kinda size of the thumb drive with fast USB 3.0 throughput, for effective deployment of deep learning capabilities. Used for real-time analysis and identification of contaminants by running machine learning and deep learning algorithms without access to the cloud.
Analysis of the captured image of the contaminated water sample from a specific place by comparing it with the clean water. If the sample water matches the properties of clean water then it declares as safe water. When the properties of sample water don’t match with the clean water then it declares as the contaminated sample. It detects the bacteria by identifying the shape, color, density, and edges of bacteria. We can able to view the images of bacteria in the sample on a computer. Suppose, the sample identified the bacteria it has is E.colli and is harmful which can cause cholera. After identification of the sample, water needs to map the water sample place in the real-time. And mark an alert flag on the specific location as the detected sample has E.coli bacteria.
The last solution includes a website built using node.js, which maps clean and contaminated water locations. The confidence level of the test sample water after training on the model using CNN is more than 95% and as high as 99% when comparing clean water with contaminated water locations.
Thus this project has the ability to extend the study further for detecting different bacteria, mineral, and other toxic substances by using CNN
Peter Ma’s Price winning Success!
His project won first place at the World Virtual GovHack, bagging a prize of $200000 Future specification of the project: Their next iteration of the first prototype phase is to have a single handy device (Internet of Things). So that anyone can personally take this device to different locations to detect the water bodies. And so people will able to avoid the contaminated water and use only safe, clean water for drinking purpose. Thus it will lower the risk of water-borne diseases.
“We have the ability to provide clean water for every man, woman and child on the Earth. What has been lacking is the collective will to accomplish this. What are we waiting for? This is the commitment we need to make to the world, now.” – Jean-Michel Cousteau
Peter Ma is a prolific contributor to the Intel® AI Academy, so you can get more of his insights there. Members also benefit from free access to the Intel AI DevCloud, which Ma used to train his model.
