Coronavirus 2019 (COVID-19) is an earth-wide pandemic which was considered in December 2019 to be initially reported. It spreads primarily across people as a result of close interactions. This virus often passes through sneezing, cough and conversation via small droplets.
In order to prevent the spread of this virus, the World Health Organization has taken steps to ensure that the people get rid of the virus, such as daily and thorough hand washing using the alcohol-based hand or with water and soap, which keeps people at least 1 meter (3 feet) away from each other and avoids public mouth and face contact. If we consider the legislation enforced by the WHO, it is clear that it is the safest way to avoid infection to stay away from others. Yet isolated people are not very realistic, at least in their everyday lives they must be able to maintain the minimum necessary social distance.
Our project is to create a risk map of the concerning area with the combination of human factors such as population, land cover type, vehicle usage, public transportation usage, etc. With considering above factors it is possible to detect areas with the risk of spreading the virus, based on the social distance.
The main two data sources which are we used for our project is satellite imagery and population data. We classified the satellite image into different land cover classes with existing map and get a probability of each class. In the other hand we have calculated probability of the different human factors depending on the population density data referred to different categories. From that we have calculated probability of social distance for the whole area. Finally with the probability of the land cover type we got the risk level for the concerning area to create the risk map.
For the processing we have used ArcGIS and ENVI software packages to develop our project. And we could generate the Python code for the processing also. But with the limitation of population data for some areas it makes some inaccuracies for the final result. As the satellite images have been disturbed with clouds and other atmospheric concerns, classification process may be not properly finished.
From a static risk map, we could build the application into a dynamic risk map that is based on real time data. A mobile application can be launched which will help recognize the risk of exposure prior to entering an area integrated with GPS technology.
As we concerned Galle and Matara Districts in Southern Province of Sri Lanka, we could see that mostly crowded urban areas has the high risk and as an overall view most of the areas don't have much risk of spreading the virus. We could apply this method for another countries to check their risk level of spreading the virus according to the population and land use factors.
From here you can view a short video on demonstration of our solution.
Find our project Resources and Relevant Codes here.