Mobile sensing in Aedes aegypti larva detection with biological feature extraction

Abstract

According to WHO, Dengue fever is the most critical and most rapidly mosquito-borne disease in the world over 50 years. Currently, the presence and detection of Aedes aegypti larvae (dengue-mosquitoes vector’s) are only quantified by human perception. In large-scale data, we need to automate the process of larvae detection and classification as much as possible. This paper introduces the new method to automate Aedes larvae. We use Culex larva for comparison. This method consists of data acquisition of recorded motion video, spatial movement patterns, and image statistical classification. The results show a significant difference between the biological movements of Aedes aegypti and Culex under the same environmental conditions. In 50 videos consisting of 25 Aedes larvae videos and 25 Culex larvae videos, the accuracy was 84%.