Milk Adulteration Detection System by Byte Crew

CLIENT INTRO:

Neurofy is a Swedish company which specializes in Veterinary Sciences. Their headquarter is in Stockholm, Sweden. Their aim is to provide Arti?cial Intelligence solutions to check the health of farm animals and products to avoid the use of chemicals which are harmful for the environment. Neurofy was founded in 2019. It consists of a team of Engineers, Scientists, University Professors and Entrepreneurs with an experience of more than 30 years in the ?eld of food industry and life science.

They work in collaboration with manufacturers and IT developers to come up with environment friendly farm solutions. In 2022, Neurofy was chosen among Top 10 innovative companies out of 700+ companies which were chosen by AstraZeneca, Business Sweden and Indian Sweden Healthcare Center. They have successfully implemented their solutions in the Scandinavian countries and now they have plans to increase their outreach to third world countries where farm practices are still primitive.

There aren’t many testing facilities in 3rd world countries and to get a product or an animal tested, the farmers have to travel long distances. They have to wait for weeks and even months to get the test results. Neurofy has plans to use Arti?cial Intelligence to provide farm solutions to the farmers which will eliminate the need to send samples to testing facilities. They will be able to diagnose their animals and control product quality with ease.

Neurofy solutions work in the following way:

Target Odor

Odor consists of chemical compounds which can be identi?ed through AI. Machine Learning Algorithms that are developed to detect different types of diseases in animals and impurities in farm products.

Manufacturing

After that Neurofy manufactures electronic devices which are capable of calculating the quantity of volatile organic compounds in an animal or animal product.

Proof of Concept

These chemicals are then analyzed in Machine Learning Algorithms which has all the information about the volatile organic compounds and their speci?c chemical compositions which chalks out a proof that whether a substance is normal or not.

Increase Accuracy

The Machine Learning Algorithms are then trained using the data they get from Neurofy devices. The more data they analyze, the more accurate they become.

EXECUTIVE SUMMARY

The purpose of this project was to provide a cheap and environment friendly solution to farmers in India to ?nd adulteration in milk. According to one of the surveys conducted in India, 68.7% milk and milk products are adulterated and are not as per the standards laid down by the Food Safety and Standards Authority of India (FSSAI). According to another report by the Indian Ministry of Science and Technology, around 89% of milk in India is adulterated.

According to the World Health Organization (WHO) if milk adulteration is not checked immediately then by 2025, 87% of Indian citizens will get serious diseases like cancer. As India has the world’s second largest population and it is close to becoming the world's largest population, the primitive ways to check adulteration in milk is very harmful for the environment which will affect the whole world. Apart from the environment friendliness, these old methods cannot cater tests in such large numbers.

The AI solution built during this project is the most feasible solution keeping in view the large number of tests needed to be done. Because the results are calculated in real time by a Machine Learning Algorithm, there will be no long queues of farmers waiting in lines for their results. Neurofy has provided pocket-sized state of the art devices to the farmers in the past which were really helpful in ?nding different diseases like Metritis, Mastitis, Ketosis and Campylobacter in animals just by analyzing their odor.

Now Neurofy wanted to get a step further to ?nd an AI solution to check the adulteration in Milk and Milk products. For that they manufactured a Neurofy Edge Device which was speci?cally made to analyze the adulteration in milk. They hired our services to make a user-friendly mobile application, an admin panel and most importantly, a Machine Learning Algorithm which is trainable, which means that it will get more accurate and fast with each test. We had faith in our capabilities because we’ve been through great challenges as a team and every time we have come out victorious. So we took it as a challenge and started assembling the most suitable people from our team for this project. The team consisted of the Design and Frontend team, Laravel (PHP) team, Python team and Quality Assurance team. All our teams worked in a perfect sync and completed the project with great success.

CHALLENGES

One of the great challenges we faced while developing this app was the unavailability of the Neurofy device. The Neurofy device is bluetooth based and we didn’t have the device physically available in our country to test the connectivity of the device with the mobile application we made.?

Another challenge for us was the writing of the machine learning algorithm. Though our machine learning and AI department are phenomenal in their work, it was not an easy task to understand all the technical aspects of the milk adulteration process and to make the Machine Learning Algorithm as accurate as possible.?

PROPOSED SOLUTIONS

We asked our client to give us a dedicated desktop machine and we connected to it remotely through AnyDesk software which allowed us to work on our client’s computer. Our technical team wasted no time in setting up the Integrated Development Environment on the client's computer so that we could directly run our android application using an emulator and connect it with the Neurofy device.

After running some test cycles, we asked our client to provide us with an android smartphone with AnyDesk access and connect it with the computer so that we could directly test the connectivity of the Neurofy Edge Device with the Neurofy Application we made. While our mobile application development team was busy in the development of the mobile application, our Python experts were busy in understanding the existing machine learning algorithms which were being used for other Neurofy products.?

After a detailed meeting with the client, our developers started writing the Machine Learning Algorithm. After getting through some road blocks, we were able to write the Machine Learning Algorithm as per the milk adulteration logic.?

We have developed the following solutions for Neurofy:

1. Mobile Application on React Native.?
2. Admin Panel in Laravel and React JS.
3. APIs for the communication between Laravel and React Native.
4. Machine Learning Algorithm.
5. Python APIs for the communication between Machine Learning Algorithm and Mobile Application.?

WORKING OF THE MILK ADULTERATION SYSTEM.?

1. User creates an account on the Neurofy Application. A farmer’s ID is generated after successful registration.
2. The user can go to the Neurofy testing center or the Mobile Testing Laboratory of Neurofy will come to the user.
3. The user will give his/her farmer’s ID to the Neurofy staff.
4. The staff member will connect the Neurofy application with the Neurofy Edge device.
5. After the connection is established, the device will take approximately 5 minutes to warm up. The warm up time depends on whether the weather is warm or cold.?
6. After the device is warmed up, the application will be ready to use.
7. The staff member will enter the farmer’s ID of the user and the user will provide a milk sample to the staff member to insert in the Neurofy Edge Device.
8. When the device and the sample is ready then the staff member will start the test. The Neurofy Edge Device will first intake clean air for 20 seconds. After 20 seconds the valve will switch and the device will intake sample air from milk for 20 seconds. The device will intake air for 20 more seconds to cool down and clean the sensors.
9. The user will have an option to retry the current test or click on the next sample to take a sample from another user.
10. After the device is done taking the sample air and clean air data, it would process this data and generate a log of raw data.
11. Our system will send this raw data to Neurofy Cloud where our Machine Learning Algorithm is housed. The Algorithm will then analyze the data and calculate whether the milk is adulterated or not.
12. Our system will get the result from the cloud and would display it in our application.?

NEUROFY EDGE DEVICES

CONCLUSION?

Instead of all the roadblocks, we were able to accomplish this task with flying colors. We could have completed this project using the old PLC server but it is our tradition to have a proactive approach and to think about the future. We went with the most scalable approach to eliminate any hindrance which might occur in the future. Our client was so happy with our work that they have already hired our expertise for the future expansion of this system. In the next phase, we will integrate the Quant Studio 7 Flex Real- Time PCR System as well as other equipment to expand the ability of the system to perform more kinds of tests.?

APPLIED TECHNOLOGIES

React Native?

React Native was chosen for the creation of android application because of its flexibility with other platforms. React Native is very good at interacting with PHP based backend frameworks.?

Python?

Python is the God of machine learning and data science. Neurofy Cloud and Neurofy Database uses AI so we had no second thoughts about using anything else for the development of the Machine Learning Algorithm. Both the Algorithm and its APIs which are used to communicate the data from the cloud to our application were written in Python.?

Laravel?

Laravel is one of our strengths. We love it because it is a development powerhouse and the best PHP framework in the market. It gives us the power to create powerful backends which enables rapid development with MVC architecture and its high end security features are required to make any modern day app secure.?

React JS?

React JS was deemed most suitable for this project because of its fast interface and seamless integration with PHP backend. At Byte Crew, we have a pretty solid front end team which is proficient in all the big three frontend technologies i.e. React, Vue and Angular JS. It works like a treat in the Laravel framework. It provides us with both the looks and the brains for a smooth and fluid experience.?

MySQL?

MySQL is one of our favorites too because of its high performance, flexibility, security, round the clock uptime and complete work flow control. MySQL is the most popular Relational Database Management System in the market. It is known for its scalability and integrability with backend languages. After some rounds of internal discussions, we finalized MySQL to be the most suitable RDBMS for this project.?

Byte Crew Technologies