Change your cover photo
Upload
Change your cover photo
budding Embedded engineer
This user account status is Approved

This user has not added any information to their profile yet.

An Efficient Multimodal AI on Edge-Based Animal Intrusion Detection and Alert System for Crop Protection
MAX78000 Feather Board

The main aim of this project is to develop a low-cost and reliable solution for unexpected crop
damages due to animal intrusion. In the last three years, there were 7500+ reported incidents of
crop-raiding by wild animals damaging around 5000 acres of land in the southern region of
India[1]. In Kerala and Tamilnadu region the elephants and wild boars are one of the main
intruders and the main crops that they destroy are : Banana, areca nut, coconut, tapioca and
sorghum [1] [2]. This problem is not just for any particular state but it is a very common issue in
a vast country like India. The farmers along the western ghats and northern region also face
various different struggles when it comes to wild animals damaging their crops [3] [4].
The existing solutions implemented currently are mostly Electric fence based which is very
harmful to the animal thus ruining our purpose to come up with an effective solution for the
farmer and protect the wildlife. There have also been other ways like using natural repellents or
Biophysical barriers but these methods often don’t work on a long term since the animals learn
to adapt and they eventually find a way to overcome these. To overcome these challenges and
ensure the safety of both the farmer and the wildlife the proposed system will be a low-cost,
scalable system designed to detect animal intrusion. The intrusion detection will be based on
laser, image and audio, the collected data is processed at the edge and the alerts are directly
sent to the user. The sensor node would be monitoring the area continuously and when there is
a wild animal intrusion this will send an alert to the concerned people which will help them make
a timely intervention and also use an ultrasonic or light based repellent. Due to animal intrusion
around 45.76 , 43.07, 31.25 % of total area under maize, wheat and paddy is prone to animal
menace [4]. Thus, we can also expect to save around that much once our solution is deployed.
The chosen hardware for this project is, Analog devices MAX78000 Feather Board. This board
comes with a built in mems microphone, low-power stereo audio CODEC and CMOS Image
sensor (camera) which would reduce the overall development cost and also the CNN
accelerator would be helpful in coming up with a model which would help us process the audio
at the edge and transmit required data.

[1] https://www.newindianexpress.com/states/tamil-nadu/2020/Sep/20/7562-cases-of-crop-raidin
g-by-wild-animals-reported-in-last-threeyears-across-tamil-nadu-2199345.html
[2] https://www.indianforester.co.in/index.php/indianforester/article/download/7324/6322
[3] https://scroll.in/article/832715/darjeelings-small-farmers-are-losing-40-of-their-crop-to-raidingwild-animals
[4] https://epubs.icar.org.in/index.php/IJAnS/article/download/124173/46928/333664# :~:text=The
proportionofareaaffected,waspronetoanimalmenace

Abinav Balachandar
Sreenath Vijayakumar
Sreehari V