Aedes Larvae Classification and Detection (ALCD) System by Using Deep Learning

Nowadays, the present of the latest technologies like Artificial Intelligence and
lenses that can capture the micro-living being like larva have been used in our
surrounding environment. Deep Learning technologies which are a subset of Artificial
Intelligence have been implemented in used for processing the image. As before this,
there is a study to detect the possible place of Aedes mosquito breeding place with the
use of Internet of Things (IoT) technologies to detect the humidity of certain places
and relate it to the possibility of Aedes mosquito breeding present. To support the
study and have verification of the place is the breeding place of Aedes mosquito, a
study to classify the larva and detect it has been proposed. The Aedes Larvae
Classification and Detection (ALCD) System by using Deep learning is a system that
uses deep learning technologies to detect the pattern of the larva and classify it
according to its type.
The proposed developed system ALCD because Malaysia is having a rapid
increase in dengue cases throughout the year. A dengue virus outbreak has started at
Malaysia as early in 1900. From 1900, the number of cases was reported to keep
increasing until recent. As on 2019 the number of cases reported was 114, 013 cases
which on the last year 2018 the number of cases reported only 80, 615 cases. The
increasing of 33, 398 cases from 2018 to 2019 has stated that Malaysia was in danger
zone and it needs an efficient method on how to combat the dengue virus outbreak.
While the death case reported from 1 January 2019 till 9 November 2019 was 158
people. The number of death and dengue cases has stated that through the years it
keeps increasing.
While there are many approaches from the government and non-government
organizations (NGOs) to help combat the dengue virus outbreak, this study is focusing
on to prevent the virus from spreading in the early stages. The life cycle of an Aedes
mosquito is starting from the egg to larva to pupa and lastly became an adult mosquito.
The early stages of Aedes mosquito that can be used to differentiate between Aedes
and Non-Aedes were at the larva stages. This study is meant to do a background study
on using the latest technology of deep learning subset of Artificial Intelligence
technology to find the pattern of the Aedes and Non-Aedes on the larva. After the pattern of the larva type is recognized, the process to classify it between the Aedes
larvae and Non-Aedes larvae can be continued for classification. Real-time
classification testing will be conducted to test the accuracy and efficiency of the ALCD
system. The portable system of the ALCD system will be developed to let the normal
people use it as it will help to fasten up the combat process of Aedes breeding site. The project development will follow the research phase methodology model
for helping in keeping this project on track. In the first phase, background studies will
be conducted for the project background and doing the theoretical studies for the
mosquito breeding cycle. In the second phase, a literature review will be conducted
for related work and past studies to see the project relevancy and have the possibility
to deliver it. The third phase, a very important phase for finding the pattern of the larva
type and to classify the larva between Aedes larvae and Non-Aedes larvae by using
Convolutional Neural Network (CNN) model from the deep learning technology. As
stated by Sanchez-Ortiz (2017), by using the CNN we can find the pattern of the larva
type and classify it according to their type. The next phase was the development of a
portable ALCD system was based on the success of phase three, this phase was meant
to develop a mobile application that is embedded with the ALCD system. The last
phase was testing the ALCD system which involved using the real specimen for
detection and classifies it according to their type.
As the study is completed, the expected result was the pattern of the larva Aedes
and Non-larva Aedes is identified. The ALCD system will able to identify Aedes
larvae. A real-time picture captured and uploaded to the server which is the ALCD
system for the analysis and classification of the larvae type. A portable ALCD system
that can be embedded in computer and mobile phone system for the ease of usability
of medical officer and normal people.