Human Activity Recognition (HAR) Computer Vision System using Deep Learning-Based Detection Corona (Covid-19) for People Mobilization

Abstract

The spread of disease will be difficult to detect in person who is active outside the home and
requires a wide active wide of motion, person who feels healthy will leave the house ignoring
normal body temperature conditions which are very dangerous, because some viruses will
mutation and easily spread to other people especially if we are in same area public, the
emergence of a virus in a person's body is the easiest to find the symptoms of body temperature
above normal such as covid 19 disease and other diseases caused by fever and shortness of
breath , the prevention solution is currently by placing several officers in other public places with
measure on temperature using thermal sensors to detect body temperature above normal between
37oC to 38oC, if the temperature indication is found unhealthy people need to prioritize and
monitoring the health conducting further tests in the form of rapid tests and swab test, the
weakness of this monitoring is the number of person passing in public areas is very lot of people
and disrupts the activities of especially those requiring high mobility need causes long and longterm checks. The inspection system able to carried out quickly, it is necessary for check and
read with computer vision technology using thermal technology have capability and accurately
for record body temperature, thermal data can read individual body conditions but unused
directly in public areas, systems machine learning will provide a visual sign that can only be seen
by the officer with the marking will make it easier for officers to identify more person's
condition and take action checking and prevention, this system has accuracy above 90% and
useful in masse and large public areas.