Drowning Detector – AI Free Code

“AI Save Life From Drowning”

In this project, we’ll build a drowning detector program. We are using YOLO object detection and can detect whether a person is drowning or it’s a normal person. Similarly, a Raspberry Pi Camera should be used for this project, and can then be put into underwater with a suitable case.

Let’s understand how this project is used to detect a drowning person. So, the initial step is to identify the person first and draw a blue rectangle box around them. After drawing the rectangular box, this project store the center position of the person from the created rectangular box. And, the program runs till 10 seconds to compare the position of the person which means to compare a person’s position is more or a person’s position is less or a person’s position is falling. If the program detects the drowning person then the blue rectangular box will now turn to red and the word ‘DROWN’ will appear on the top of the person. Similarly, if the program detects the center position of the person is above water, then the term ‘DROWN’ would be omitted by the program as a person are only standing in the pond. 

During the research, what we have found is that if the person or someone drowns, they prefer to remain in a vertical position and attempt and keep their head above the water level. This may be examined by observing the ratio of height-to-width of the triangle to decide horizontal and vertical person of someone to find they are swimming or drowning. 

NOTE: If you are interested in YOLO object detection. (view link)

Step involved to run this drowning detector program: 

  • 1. Download the repository and unzip the file.
  • 2. Install all required libraries
  • 3. Open DrownDetect.py and here you go.
  • 4. Enjoy drowning detector program
  • 5. Finally, help to make lifeguards in the swimming pool/River.

!! DOWNLOAD CODE NOW !!

Credit: Nico

4 thoughts on “Drowning Detector – AI Free Code

Leave a Reply

Your email address will not be published. Required fields are marked *