Abstract:
"
Anyone in this world has heard of the disasters happen due to deadly cyclones. Some may have
seen them and even some may have actually faced such situations. Thousands of people loss their
loved ones, loss their properties annually due to this deadly disaster.
Even we don’t have the ability to stop such disasters, but we can early predict them and place the
people in safe locations. Where we could definitely save lives. For detecting such disasters, people
have understood the environmental distortions happen before such disasters and use the
information gathered to predict and figure out. But with the upscale of technology, people got the
ability to view the environmental changes being above the ground level using arial vehicles. This
what we call remote sensing. In the present, satellites were used to broadcast live video data to the
ground for further processing. But most of the processing is done with human participation, which
could definitely reduce the efficiency, accuracy and speed of processing.
In this Project, the author has come up with a computer-based solution which could increase the
potential of determining a cyclone before it hits the human society. All the live data broadcasted
by the satellite are gone through an image processing system and identify each frame’s availability
of a cyclone. Before the detection, the model has been trained using deep learning techniques."
