Bird predation is a major problem in aquaculture. A novel method for dispersing
birds is the use of a vehicle that can sense and chase birds. Image recognition software
can improve their efficiency in chasing birds. Three recognition techniques were tested to
identify birds 1) image morphology 2) artificial neural networks, and 3) template
matching have been tested. A study was conducted on three species of birds 1) pelicans,
2) egrets, and 3) cormorants. Images are divided into three types: 1) Type 1, 2) Type 2 and
3) Separate photo type 3 from other types based on difficulty. this
The types were clear, moderately clear, and unclear, respectively. Image morphology arose
Correct classification rate (CCR) from 57.1% to 97.7%, 73.0% to 100%, 46.1% to 95.5%
Pelican, cormorant, and heron images before the size threshold, respectively. Or
The artificial neural network model is a 100 ° R test type 1 image and its
The success of the classification during the test was between 63.5% and 70.0%, and between 57.1% and 67.7%.
Type 2 or Type 3 images. The template matching algorithm was successful.
90%, 80%, 60% classification of pelican type 1, type 2, and type 3 images
Heron. This technique recognized 80%, 91.7%, and 80% of Type 1, Type 2, and Type 3
images of cormorants. We developed a real time recognition algorithm that could capture images from a
camera, process them, and send output to the autonomous boat in regular intervals of
time. Future research will focus on testing the recognition algorithms in natural or
aquacultural settings on autonomous boats.
suntech software solutions provides best switch to on bird specie detection raspberry pi project with complete description and project report, and complete on call guidance related to the project order now.
Component List
- raspberry pi
- webCam
- Powersupply
- wires and connector
Features
- detectects bird specie which helps photographers and geology parks to understand and detect the different birds
