Abstract
Mosquitos pose a severe health hazard by transmitting dangerous vector borne diseases, currently existing mosquito prevention and detection techniques are labour extensive and are not highly accurate. The DragonFly robot platform used in the project captures mosquitos using different methods to attract mosquitos and trap them safely, though it is an autonomous robot, it lacks the ability to perceive environment in-terms of mosquito diversity. This research was focused on detection and tracking and lacks the ability to map the diversity of mosquitos to efficiently tackle the situation. The State of The Art Object Detection model YOLO-NAS, was used to identify the hotspots which mosquitos breed and accumulate and also the algorithm was trained to identify the species of captured mosquitos. Finally, the results were plotted on the map that was generated by the existing ROS SLAM (Simultaneous Location and Mapping), to investigate on the distribution and diversity of mosquitos. Various challenges and the methods used to tackle them, are mentioned step-by-step in the thesis. The drawbacks and limitations of the research work is also presented as a conclusion to the research.