High-Level Project Summary
In this challenge, I have worked on the satellite images and the behaviour of human navigation for Measuring the Value of Earth Observations based on the Earth observations data to find missing persons. The problem of "missing people" happens each month all over the world, especially in the areas of mountains, seas or deserts. Using multispectral satellite, geographic profiling and machine learning technology to reduce the time and area for searching and predict the potential areas of missing people.
Link to Project "Demo"
Link to Final Project
Detailed Project Description
I created an app that helps residents to submit information on a missing traveller and the information will submit to the police right away for seeking the one. At the simple platform, the users firstly provide personal information such as name, age, address and national identity ID. Then more detail needs to be supported including the latest characteristic and position where the person was last seen from the app user. In this step, the user provides both frequent places and the latest route that the missing person visited.
The app will search the position of the missing person based on their GPS information from their car or devices. In the simple version, I admit the missing one has a GPS device that can be tracked by satellite. The next case is losing the GPS signal. In this stage, I use Machine learning techniques (Yolo3) to train data from personal information and satellite images. The result will show the position of missing persons on the map.
Space Agency Data
I used a lot of data from space agencies by collecting data of pictures and videos from multiple sources such as Google search, open sources, NASA's datastores. I also used NASA API, Google Maps API while creating a simple app within React technology.
Hackathon Journey
Every year, thousands of people are missing all over the world, especially in the deserts, mountains or seas where people prefer to travel or explore nature. The government and other organizations are spending lots of time and money seeking missing persons. All the information are based on the family, friends' story or CCTV data. However, it is still a challenge for the areas far from urban with not much detailed information as listed above. Unfortunately, data on missing people can be scarce and unreliable. Then the space technology and machine learning algorithm should combine to support solving this problem.
From the Space App challenge, I understood more about how the importance of Space technology impact the Earth such as how NASA uses satellites to monitor the Earth, understand the GIS terminology and many challenging that are still needed to solve related to human life, agriculture, climate change, forestry and urban.
If we can install the app on the phone, smartwatch or other devices within GPS technology, the first advantage of tracking technology is to record the position data which help to find the latest position of the owner in the case they are missing. Collecting other data such as terrain, road routes, and multispectral images for training data is to help machine learning models predict the area of missing people.
I was thinking about how to collect data from multiple sources, which techniques can be used to train models and predict the area of missing people and how to create a potential app. I did literature reviews on each of the steps to find suitable solutions. Although the app is still building, the proposal is the potential for "search and rescue" for missing people cases. The future apps can be extended to other applications such as wildlife monitoring, mining vehicles operations or even criminal investigation.
References
VSCode, NASA API, Google Maps, JS, React, Drive
Tags
#Human, #Machine Learning, #Satellite Images
Global Judging
This project has been submitted for consideration during the Judging process.