Birds protection with oneM2M (by Saeboda team)

Story

In Korea, various accidents such as annual power outages, sanitation problems, crop management problems, and landscape deterioration continue to occur.

As evidence of that, 100 million won a year is being used for damage from new electric poles. Therefore, to prevent damage to electric poles caused by such birds, we designed the 'Saeboda' service.

Build Instructions

The hardware consists of two sub motors, one buzzer, one Raspberry Pi, and a camera module.

• Submotor for operating the atomizer
• Submotor for rotating the atomizer
• Camera module for bird recognition
• buzzer for noise

and additionally a solar generator for continuous charging

And the modeled hardware was output.

Ask for the right pedestal to wrap around the power pole, and use it by inserting the house on the left in a detachable form.

Code

For this project, we first had to create a bird-aware AI, for which we used Kaggle's 200 Bird Species with 11, 788 Images (https://www.kaggle.com/veeralakrishna/200-bird-species-with-11788-images) dataset. I used it.

All of the existing images were of different sizes, so they were unified and changed from color to black and white to make the model lighter.

import numpy as np 
from PIL import ImageGrab 
from PIL import Image 
import cv2 

screen = Image.open(source) 
img_resize = cv2.resize(screen, (518,518) 
img = Image.fromarray(img_resize).convert('L')

After that, the data was divided into 70% training data and 30% test data, and the Yolo model downloaded through github was trained.

Finally I put all the parts in the box.

Execution Scenario

1. find bird

2. noise generation

3. watering with a sprayer

4. Record where the bird was found in Mobius

But I failed to spray water and link with Mobius platform.