Kambio Verde

Demo: https://youtu.be/hXC8iSLEKuM

Proyecto: https://sea-life-kv.web.app

Have you ever wondered where the plastic we throw away ends up? Currently, 150 million* metric tons of plastic waste are in our oceans and each year, an additional 8 million tons are added, affecting the marine’s fauna and flora. Therefore, using Artificial Intelligence and Machine Learning, we created SeaLive, the "Protobasura" that creates life; a comprehensive model accessible and understandable to the general public. With it, you can identify the route of plastic waste to locate, classify and quantify it according to its level of toxicity and, in order to manage their treatment with private or public entities, contribute to the conservation of the ocean, which is responsible for the existence of all living beings on our planet.

Have you ever wondered where the plastic we throw away ends up? Currently, 150 million* metric tons of plastic waste are in our oceans and each year, an additional 8 million tons are added, affecting the marine’s fauna and flora. Therefore, using Artificial Intelligence and Machine Learning, we created SeaLive, the "Protobasura" that creates life; a comprehensive model accessible and understandable to the general public. With it, you can identify the route of plastic waste to locate, classify and quantify it according to its level of toxicity and, in order to manage their treatment with private or public entities, contribute to the conservation of the ocean, which is responsible for the existence of all living beings on our planet.

SeaLive, the "protobasura" that creates life, is an integral model that seeks to contribute sustainably to the preservation of the oceans, rivers and coasts; taking advantage of existing technology to identify the route of plastic waste in order to locate and identify it; then classify it to quantify it and thus manage it’s impact on the environment; making the population aware of how to classify their waste and integrate it into environmental programs that promote the preservation and care of our oceans.

In Peru, according to data from the Ministry of the Environment based on a beach cleanup campaign along the coast, "in the last quarter of 2010, 29,910 metric tons of solid waste was collected. Of the total, most (46.5%)* corresponded to plastic material". Among the plastic materials found were "plastic bags, bottles, cups and plastic cutlery". These are definitely alarming figures that define a path to incorporate actions that can prevent this waste from reaching our oceans in advance.

The three pillars, identification, classification and awareness of plastic waste, determine a model that enriches each other, thanks to the information collected at each stage. Our solution starts with protobasuras* (plastic waste) that is thrown into the sea, rivers, among others, and that through its GPS defines the route it takes through the ocean. This allows us to identify the route or routes that the waste can use to reach the large waste banks. These routes will be demarcated and used as a "Waze" (application used to choose the best route and avoid traffic) in order to build the marine pathways through which plastic waste travels to create new waste banks and identify the pathways of existing waste banks that are not demarcated yet. When it identifies the route and finds waste in the ocean, it issues an alert through its GPS and a Drone that is strategically located along the routes - predetermined by NASA satellites, taking advantage of CYGNSS* data -, takes a picture of the focus of the waste, sends the information and, through a Machine Learning model of "image identification", classifies the plastic waste, defining its degree of risk to the environment and saving the identified route. This model classifies the garbage by defining a record of waste in the routes taken by the protobasura* so that the government, scientists and people who want to access our solution can know the type of plastic waste and define environmental actions that can manage the cleanup and/or use of this waste. As we mentioned at the beginning, awareness is very relevant in our model because the waste does not reach the oceans by itself; the human being is the antagonist in this story and his role must change to be the action hero or superhero actively collaborating in the knowledge and preservation of plastic waste.

Our Benefits

·      Identify the routes of plastic debris, achieving a definition of routes along the ocean that allows to determine preventive actions by them

·      Classify waste (machine learning) derived from plastic in order to define cleanup and environmental management projects.

·      Society should be the manager of change in the knowledge of how to dispose of plastic waste so that it can be used for other purposes.

·      Scientists, environmentalists and the government should know the degree of contamination by plastic waste to generate real initiatives to preserve the environment. 

·      Our model is the initial basis for future experiments where scientists and students who want to participate in a collaborative environment from different universities around the world can use it as a basis to improve it to make it sustainable over time.

Through SeaLive, the "protobasura" that creates life, we want to make information available and easily accessible to all stakeholders and ordinary people, so that they can easily understand how not disposing of plastic waste correctly can affect their environment. We also hope that environmentalists, the government and other entities can build clear policies that preserve the environment. The principle of making the information available makes it easily accessible and therefore anyone can know it, understand it, and start generating actions such as feeding our image bank of plastic waste that will also be used to improve our machine learning model.

The tools used for the integral model of SeaLive "protobasura" that creates Life were:

• Modelo de Machine Learning: Red neuronal YOLO. Librerías utilizadas: tensorflow versión 1.13.2 / keras versión 2.0.8 / imgaug versión 0.2.5 / opencv / h5py / tqdm / imutils / PyQt5 / labelImg
• Frontend: Lenguaje JavaScript 
• Framework: Vue.Js / Librería UI: Vuetify
• Prototipo: Figma.
• Apis Externas: Maps de Google Cloud
• Hardware: Laptop Lenovo YOGA 510-15isk

Software: Windows 10 / Procesador: Intel(R) Core(TM) i7-6500U CPU @ 2.50GHz 2.60 GHz / RAM instalada: 8.00 GB (7.86 GB usable)

*Fuente: https://www.wwf.org.pe/

*Protobasura: Technologically engineered debris created from dense, plastic waste material that is thrown into the ocean).

*CYGNSS: The new technique is based on data from NASA's Cyclone Global Navigation Satellite System (CYGNSS), a constellation of eight small satellites that measures wind speed over the Earth's oceans and provides information on hurricane strength. CYGNSS also uses radar to measure ocean roughness, which is affected by several factors, including wind speed and debris floating on the water.

*Figures from the Peruvian Ministry of the Environment's guide to environmental education in coastal marine areas.

• Plastic marine debris is one of the most dangerous global threats, due to its longevity in the marine environment, the physical and chemical hazards it presents to marine and bird life, causing economic losses to the fishing and maritime industries and degrading the quality of life in coastal communities.

• Understanding the problem and knowing in detail the actions that have been taken in the reduction of marine debris was the starting point to address our challenge; SeaLive takes as a fundamental part of its understanding the "quick data on marine rights" to know how the initiatives of the marine debris program mission have generated the removal of more than "22,500 metric tons of waste", the positive impact on the proper management of marine debris to generate "tourism and recreation" in the cities and how more than 800 marine species have been affected by this problem.

• Knowing the experiment conducted by the National Oceanic and Atmospheric Administration (NOAA) we were able to understand how this model identified the microparticles of plastic waste by locating "garbage patches" in the ocean and how through the "buoys" they were able to collect a large amount of data that contribute to a model where large amounts of microparticles were visualized. We wanted to apply this knowledge in the first phase or pillar of SeaLive because it seeks to identify through "protobasuras" the route that carries out these wastes so that it can be used for the analysis and definition of environmental actions of impact on the marine ecosystem.

• The experiment conducted by the "AIRSAR mission" helped us define how the drones that we will use in our model will contribute to identify through this route of the ocean plastic debris, classifying it through machine learning models "convolutional neural network" for the detection of images and then classify it according to their risk so that this information is easily accessible and understandable for people who want to use it.

·      Marine Debris

·      SVS: Garbage Patch Visualization Experiment

·      AIRSAR JPL/NASA, Welcome !

·      Marine debris impacts

·      Toxicological tareas of plastics

For the KambioVerde team to participate in the "Space Apps Challenge 2021" it has been a great adventure, it is the first time we participate in this "hackathon" and it has been really enriching. Each one of us has participated in other events like this, but the adrenaline that this one produces by having the support of NASA definitely generates an even greater commitment. All the information described for the challenges and the reference data to address the different initiatives confirm how these events where "you put creativity to fly" become solutions that can help to sustainably change the planet. With the challenge we chose: "AI / ML APPROVEMENT FOR PLASTIC MARINE WASTE" we definitely learned how research and consider the information from governmental and non-governmental entities in order to generate great knowledge of the grand problem that affects our planet and that is still unknown but easily accessible to everyone. We understood that to change the world we must start with the change ourselves, and that if you start at home, giving a different use to your own plastics, you start the contribution to the planet.

Our inspiration to choose the care of the seas against plastic waste was focused on knowing that each peruvian uses an average of "30 kilograms of plastic per year" and that "46% of plastic waste in Lima and Callao are single-use". Situated in this problem and knowing that in some sectors of the city there are large accumulations of garbage affecting the population residing there (respiratory diseases, stomach and sometimes even death) this is transposed to the large banks of marine debris that affect the marine fauna and flora, making us aware that the "coastal marine areas of our country are some of the richest in the world in terms of biomass and diversity".

We focus on developing an integral model where each phase of this process delivers value and meaning to our stakeholders. Technology as a means to integrate the entire model into three important pillars: Identification, Classification and Awareness, which are all vital for the model to be scalable.

Identification: The "protobasura" is the initial input for the whole model to articulate, it is designed under the specifications of a normal plastic waste but it carries with it a built-in GPS which allows this "prototrash" to travel as normal waste would, identifying the course that the waste takes to know the different routes of plastic waste and identify other unknown waste banks. In real time we can see the location and save it in the database that allows us to design the route of the plastic waste.

Classification: "Protobasura" launches an alert to different drones located at strategic points. It works taking into account the research of Nasa scientists where they use satellites that locate the microparticles of the debris to locate all the plastic waste. From these locations, drones locate the "protobasura" taking pictures and sending them in real time to the database that feed our Machine Learning model and through a neural network that classifies the plastic waste that finds the "protobasura", making sure that it can be a new waste bank or the waste already existing in a previously identified bank (https://www.nasa.gov/feature/esnt2021/scientists-use-nasa-satellite-data-to-track-ocean-microplastics-from-space). The classification of the waste is through the neural network that through image detection classifies the plastic waste with its risk level.

Awareness: The population is immersed in this model as a protagonist from two perspectives. Since through the same platform ordinary people can go referencing waste, you find on the beaches through photos to classify them (image bank for improvement of the Deep learning model) and from the more academic management where students and scientists who want to join the cause can use our model as a basis for future and better research in a collaborative process of co-creation.

The problem of classification and quantification of plastic waste is a great challenge, many studies have been conducted that allow us to recognize that we are facing a problem that grows every day and generates major catastrophes in our oceans. That is why SeaLive contributes as an integral model that uses existing technology with the help of artificial intelligence and machine learning to create a sustainable process that allows to collaboratively improve it and make it sustainable over time.

The development of this model started with the necessary research of the sources of information that define the route that we followed to focus our initiative.

Image made for the understanding of our challenge.

After doing the research and understanding our challenge, we delimited through a prototype how we were going to develop the model https://www.figma.com/file/1wxEv7gfncck1tHh3qVmUk/Hackathon-Nasa?node-id=0%3A1.We defined the three pillars that were to come to life in the development of the IDENTIFICATION of plastic waste through the marine "waze" called "protobasura".

Image made for the understanding of our challenge.

After doing the research and understanding our challenge, we delimited through a prototype how we were going to develop the model https://www.figma.com/file/1wxEv7gfncck1tHh3qVmUk/Hackathon-Nasa?node-id=0%3A1.We defined the three pillars that were to come to life in the development of the IDENTIFICATION of plastic waste through the marine "waze" called "protobasura".

Image made for the understanding of our challenge.

We then focused on the CLASSIFICATION of the debris with deep learning through a neural network for image identification that classifies and quantifies the debris.

Image made for the understanding of our challenge.

Finally, the most important pillar is awareness raising, where society contributes in two ways; from the ordinary citizen who can take pictures of plastic waste and leave them in the image bank of the tool to improve the models and with scientists and universities that want to take this model as a basis to improve it and make it sustainable. Knowledge is information and information should be freely and easily accessible.

Image made for the understanding of our challenge.

If we had to make a simile of what we experienced in the "hackathon" we would associate it to an amusement park, where there are games that are fun and entertaining, others that are not so fun and others where you release all the adrenaline. We went through very comfortable moments where we observed that we were all advancing in a surprising way, other moments where frustration was part of the moment due to fatigue and moments where the energy was at 100% releasing dopamine because we were materializing what we had planned in Figma. Communication was essential, the Check-In in the morning was super necessary, we would then meet again in the afternoon and the check out of the day contributed to all of us arriving with the tasks done. We divided the group into subgroups by tasks that helped the whole project to be elaborated.

We want to thank NASA for these spaces that generate new geographies, recognizing talent in other latitudes. We also want to thank all the groups that participate in this type of activities that involve actions that can change the world.

Source: Figures Ministry of the Environment: World and Peruvian Figures | Less Plastic, More Life

·      Universidad de Barcelona. (2021). MARLIT, una aplicación basada en la inteligencia artificial para estudiar los macrorresiduos marinos flotantes. Obtenido de https://www.ub.edu/web/ub/es/menu_eines/noticies/2021/02/003.html

·      National Geographic Eapaña. (2021). Inteligencia artificial para detectar la basura marina. Obtenido de https://www.nationalgeographic.com.es/mundo-ng/nuevas-tecnicas-deteccion-y-contabilizacion-basura-que-flota-mar_16717.

·      Ministerio del Ambiente. (2020). Cifras del mundo y el Perú. Obtenido de https://www.minam.gob.pe/menos-plastico-mas-vida/cifras-del-mundo-y-el-peru/.

·      Ministerio del Ambiente. (2015). Guia de educación ambiental en zona marino costeras. Obtenido de https://www.minam.gob.pe/educacion/wp-content/uploads/sites/20/2015/02/1.0-GUIA-PARA-CAMPA%C3%91AS-modelo.pdf

·      WWF. (2021). El problema del plástico en la naturaleza y como puedes ayudar. Obtenido de https://www.worldwildlife.org/descubre-wwf/historias/el-problema-del-plastico-en-la-naturaleza-y-como-puedes-ayudar.

·nasa ·challenge ·SeaLive ·protobasura ·ocean ·A ·ML ·spaceappschallenge ·Sea ·enviroment

This project has been submitted for consideration during the Judging process.