Widur - Wild Animal Intrusion Detection and Classification using Docimetry
G H Giri
Kazhakkoottam, Kerala
- 0 Collaborators
A system which provides an alert message by recognising the wild animal and the total number of animals which enter a tribal village from the forest to reduce the increasing death due to wild animals in places near to a forest using AI and Dosimetry. ...learn more
Project status: Under Development
RealSense™, Networking, Internet of Things, Artificial Intelligence
Groups
Student Developers for AI
Intel Technologies
AI DevCloud / Xeon,
Intel Python,
OpenVINO,
Other
Overview / Usage
There is always news about death/casualties due to wild animal encounter by humans in the tribal villages and also destruction of agriculture due to the interference of wild animals from forest. This system provides alert to the forest officials and locals of a tribal village about the wild animal whenever it enters the human occupied place from the forest and thereby reducing casualties.
Methodology / Approach
The data set of attenuation of a radio signal between transmitter and receiver when humans and pet animals pass through is collected and trained to make a model which recognizes humans and pet animals once it is deployed. So whenever another considerably big living creature passes between the transmitter and receiver, a trigger signal is made which captures a picture and finds the type of wild animal through image processing by existing techniques.
Since this system send only the sensor values consistently to the cloud for processing, it reduces the data consumption, power consumption and could be deployed in regions where there is no considerable strength of mobile network.
Technologies Used
The technologies to be used are AI DevCloud, Artificial Intelligence, Image Processing, Radio wave generation and detection, Sci Kit Learn, Tensor Flow.