AirView

Automated monitoring as mitigation measures against air pollution using AI.

Environmental pollution has become a critical issue due to increased demand for products and development. It is estimated that about 50% of all pollution is as a result of industrial and manufacturing activities

Governments across the globe have been taking various mitigating measures to keep a check on pollution levels by implementing various committees and plans. However, since large number of personnel and equipment are involved in the sampling, data reporting, etc., under these measures, it increases the probability of variation and personnel biases reflecting in the data, hence it is pertinent to mention that these data be treated as indicative rather than absolute. 

Similarly, various environmental monitoring measures are simply indicative and a hassle in current abating monitoring practices which is fuel to this alarmingly rising levels of pollution and should be developed with the evolving technologies. 

To monitor and prevent reckless practices, and alert government authorities simultaneously, AirView is the solution. It automates monitoring and regulation process by detecting the pollution released by the factory and drawing a heat map on where the pollution has been released and if the pollution level rises beyond the specified limit, the model will immediately alert the authority. 

Due to deforestation and industrialization, air quality is constantly degrading. Based on the increased counts of factories and manufacturing processes, both large and small scale, gaseous emissions have continued to compound. This makes industrial pollution one of the main causes of air pollution. With unplanned and rapid industrial development, all countries generate industrial waste quickly. Most of these factories release chimney wastes, particulate matter and gaseous pollutants directly into the air illegally, and these industrial pollutants are responsible for thousands of illnesses and premature deaths across the globe. Industrial pollution not only encompasses damage to structures and buildings but increased risks of different occupational hazards as well.

For instance, one of the worst industrial disasters of all time that took place in Bhopal, India, in 1984 claimed the lives of more than 8,000 people and the effects were still being felt more than two decades later. This means, industrial air pollution may not manifest immediately but takes several years.

The time has come to have a monitoring system that can control this scenario of industries polluting the environment. This has been there in the media but under the tag of dubious covering and that is why this entire concept is very crucial in the domain of industries being the major chunk for air pollution in a lot of countries.

The existing monitoring and mitigating measures are obsolete which derive a rather indicative result than absolute and this lack of strict monitoring results in the reckless and continued practice of industrial pollution. So, smarter and more efficient monitoring technology is the need of the hour.

Air pollution kills an estimated seven million people worldwide every year. WHO data shows that almost all of the global population (99%) breathe air that exceeds WHO guideline limits containing high levels of pollutants, with low- and middle-income countries suffering from the highest exposures. Ambient air pollution accounts for an estimated 4.2 million deaths per year due to stroke, heart disease, lung cancer, and chronic respiratory diseases. Around 91% of the world's population lives in places where air quality levels exceed WHO limits.

Therefore the solution will serve the entire mankind on the planet that has been affected by air pollution directly or indirectly. The targeted audience will be divided into two major chunks, the primary ones being the factories themselves and the secondary ones being the indirect population of the region that is getting affected by the air pollution.

The targeted population may vary from the factories that a peculiar government body wants to monitor to a factory that wants to work on the ESG goals of their company.

The quantification of the current market is approximately 48% of the total industries in our country India that remain unchecked due to no or less help in this domain.

The current demographics are global as this is a global issue that needs to be solved.

We conducted various surveys and user research to conclude:

(The current user research was based on a number of surveys that was conducted in an industrial area and a government sector to quantify the overall output from the research)

The current research-backed our problem statement about the industries finding a loophole in the current pollution observing rules and practicing reckless disposal system owing to the lack of strict and compulsive mitigating actions.

We've been working on a  full fledged user research to back our solution and to able to identify the needs of the population and with the statistical analysis we can say that our novel solution is something that can be a crucial game changer in terms of controlling and checking on various location and factories for the air pollution related issues.

To the best of our knowledge, AirView will be the first implementation of explainable AI on pollution detection and plotting heatmaps that can be material specific with the help of computer-aided engineering software integration.

We have developed software proposition, with ML model using explainable AI with smart techniques to generate heatmaps, of AirView and are looking forward to collaborate with domain experts to implement this conceptual and software proposition further with planned hardware modules and software architecture with a very promising business model.

Currently, we have been working on its implementation and would love to get some help from this platform for our further goals that we would love to achieve. We have been working on the real-life implementation , a user research in this domain that has backed our concept and we've been establishing a network of Sustainable Goal Policy makers to work on the policy issues via including this system as a regulatory device for various companies. 

Our proposed solution involves the following approach and technologies.

1) We have used Artificial Intelligence to detect the polllution released by the factory.

2) We experimented with several different AI models like ResNet50, DenseNet169, Inception v3. We got better results using DenseNet169.

3) Due to the lack of dataset, we have collected our dataset from opensource like google images etc.

4) We collected images of factories releasing pollution as well as images of pleasent nature.

5) We trained our model on our dataset and we have applied the booming explainable AI technique which produces the heatmap based on where it identifies the carbon emmision.

6) We used the most popular techniques called Grad-CAM, Grad-CAM++, LIME, Score-CAM.

7) These explainable AI approaches take the output prediction whether there is pollution or not and based on the output.  This approach will use the last layer of convolutional network to calculate the weighted averages of feature maps which are differentiated with respect to output class, this will later produce a heatmap and will be superimposed on the original image.

8) We have made the model very lite using TensorFlow-lite so that it be used in smaller devices and reduced cost.

9) We have implemented in such a way that it works live as well as on videos and Alert the government authorities as soon as it is detected.

10) Technologies Used:- OpenCV, TensorFlow, TensorFlow-lite, Explainable AI, Optimization techniques used:- Quantization, weight pruning etc. to optimize for better performance.

The targeted audience may vary from the factories that a peculiar government body wants to keep an eye on to a factory that wants to work on the ESG goals of their company. 

The current number of people that we're serving right now includes 30 members from the Indian Institute Of Technology's department and we're expecting this model to be a huge success and till the next year, we're hoping to serve this solution to a lot of government policymakers and a lot of factories that will make it to fifty thousand users in numbers.

The quantification of the current market is approximately 48% of the total industries in India that remain unchecked due to no or less help in this domain.

The current user research was based on a number of surveys that were conducted in an industrial area and a government sector to quantify the overall output from the research and support the number mentioned earlier.

We have a modular plan.

We're working on a revenue generation model that can be hybridized into every possible aspect. The earlier model that we'll be using will be the MoFuBoFuToFu model in which we've been working on the top cubicles of the models and are already ready now. 

We would love to invest our time in marketing and helping to provide a better base for the investors to invest in this solution so that it can be used all around the globe. The marketing part is our priority as the user research that we conducted backs up our novel concept as something that will be highly beneficial for checking on air pollution and its problems and therefore we would love to share the model with our stakeholders for further implementations.

Our product cum software is novel and has a real-life UI that helps in lowering the air pollution this is something that will have a major advantage in this domain. For measuring the progress we have a modular plan and the datelines and with the help of the datelines and the specific achievements, we're measuring our progress.

The UN SDG's that are directly in our solutions are as follows: 

SDG3 , as our novel solution is directly helping in keeping a check on the air pollution this is directly responsible for achieving good health and well-being.

SDG 11, the novel solution that we propose is directly helping in attaining this peculiar SDG by making the cities and the communities more sustainable directly and indirectly.

SDG 9, as our novel innovative concept achieves this sustainable development in the infrastructure and provides us the opportunity for more innovation in this domain.

SDG 13, the core of our project is dedicated to achieving this SDG and provides a brilliant base for the problem of climate change to be addressed.

There might be some visible and some nonvisible barriers for us to accomplish but we're sure that we'll be overcoming those to accomplish our goals. The prime barrier that we have is funding. The funds play a major role in the development of the overall solution and this is something that we'd love to get some help with. 

Apart from this, we have been halted at a certain point in the marketing domain in terms of having less networking to make progress in sharing the product with more and more consumers and stakeholders. 

From research and project management, developing business strategies to ML and AI Developers, we worked on developing this project with the conviction in exploring and innovating modern solutions to evolving problems across the globe. We are a group of undergrad technology enthusiasts and skilled in different domains through various national and international project works.

With a varied set of prospects on the table, we have developed this project with a very detailed yet analytical research, a very promising business model, a very vivid technical setup with design and implementation of explainable AI on pollution detection and plotting heatmaps for the first time.

We have not partnered with any other organization or community yet, but are looking forward to collaborate with suitable platforms to expand our horizons.

NA

The two most widely-cited estimates attribute around 7 million deaths per year to air pollution and out of that 40% of the total deaths were of women and girls that got stuck in the loop of never resolving air pollution issues.

Our novel solution not only aims to improve the quality of life for women and girls but also helps them to live a life via saving them from the contagious disease that comes from the air pollution accounted to the factories in the region.