Smart Snake Identification System using Video Processing

There is a common fact in the world that all snakes are venomous and dangerous to humans. This is due to a lack of awareness about snake species among the general public. However, based on the literature, the reality is that only 41 out of 108 reptiles are venomous and dangerous to humans. The challenge is specifically identifying the various snake species instead of considering all of them to be venomous. With the population growth frequency of snakebites reported in hospitals have arisen because of people attempt to harm snakes. This paper proposes an approach to help people in identifying snakes in panic situations using video processing, and then alert the nearest rescuer teams. This study has been carried in Sri Lanka, with a contribution to the scientific world to save both snakes and humans. The implemented system comprises of a mobile application with features including offline real-time snake identification, online real-time snake identification using video processing, manual snake detection, and alert nearest rescuers. The obtained results indicate that this application has an offline snake identification accuracy of 75%-80% and an online snake identification accuracy of 90%- 95%.