Preliminary Research and Documentation

Modules

The complete project comprises of the following modules:

• Real time object tracking system developed on AI/ML platforms using high resolution cameras
• Intruder alert system

 

Real time Object Tracking System

TheA real time object tracking system is designed to identify, tag and monitor people.people and objects and store their image information in a database. Software applications can process this data and create workflows that help in securing seapods, and tracking boats, yachts and even.

Boat Identification

High resolution cameras, installed at strategic points like beaches, buoys and seapods continuously scan peopleand monitor the water surface for ships, boats and tagyachts. them.Underwater Acameras are installed on buoys, seapod bases, or mounted on rocks to monitor underwater activity. Data and images from both sources are collated and a unique ID is generated for each personboat nearthat passes by these strategic locations. The unique ID along with the waterimage surface or in the water. Thiscode is stored in a database. Technologies like TensorFlow or YOLO (You Only Look Once) use AI algorithmsand ML to calculateanalyze the heightship ofand a personlog whenis hecreated every time the ship or sheboat inpasses by.

Optionally, a unique display code, QR code or AR code is generated and assigned for each boat by legal authorities. These are displayed on the beachship orand oncan thebe seapod.scanned and identified by cameras and software from far away.

Figure 1 - Sample image to illustrate real time object tracking and tagging on a beachboats using AI.

When an unidentified boat approaches a sea pod, the software triggers an alarm and notifies the residents, security personnel and legal authorities.

 

Intruder Identification

Intruder identification and theft prevention comprises of the following modules:

• Motion sensors
• Facial recognition system
• Alarm system

All 3 systems are linked to the software console that monitors resident and seapod safety.

Motion Sensors

Motion sensors or movement sensors are electronic devices that detect and measure movement. It uses ultrasonic sensor technology and emit sound waves to detect presence of objects. Motion sensors are installed in different areas of the seapod. When residents are away, these motion sensors scan the covered area for intruders. When motion is detected, the sensor triggers an alarm and sends a notification to the resident and local authorities.

 

Facial Recognition System

High resolution cameras, installed at strategic points inside and outside seapods, buoys and boats continuously scan and monitor the area for intruders. Underwater cameras are installed on buoys, seapod bases, or mounted on rocks to monitor intruders trying to gain underwater access to seapods.

Intruder identification software allows the 

Images and details of seapod owners and residents are pre-entered in the software and stored in the database. Residents can add as many guests, friends and families 

Data and images from both sources are collated and a unique ID is generated for each boat that passes by these strategic locations. The unique ID along with the image code is stored in a database. Technologies like TensorFlow or YOLO (You Only Look Once) use AI and ML to analyze the ship and a log is created every time the ship or boat passes by.

 

When this person enters into the water, the AI module monitors the person position, section of body that is in water. Depending on the section of body that is above water, the algorithm tags them with a safety indicator.

For example, the following indicators are displayed on the console:

• GREEN - Person is on sand, or a hard surface, is stationary or moving
• GREEN - More than 50% of the body is visible, and is stationary
• GREEN - More than 50% of the body is visible and is moving
• YELLOW - Between 10% and 40% of the body is visible and stationary
• YELLOW - Between 10% and 4% of the body is visible and stationary
• RED - Less than 10% of the person is not visible for more than 30 seconds.

Each of these sample indicators will change in real time depending on the swimmer's images that are visible to the high resolution cameras. A log is written to the system and the database is constantly updated.

If a swimmer does not surface for more than 30 seconds, the software sends a signal and deploys the the unmanned underwater vehicle (rescue bot). Alternatively, if the monitoring lifeguard sees a red indicator on a particular person, he or she can manually deploy the UUV for rescue.

 

Wrist Tag

A GPS or RFID wrist tag is to be worn by swimmers when going into water. The GPS or RFID unit gives the exact position of the person wearing it. The wrist tag also monitors the pulse or heartbeat of the swimmer and relays the same to the console. If there is an irregularity in the pulse, and the person is in water, the person is tagged as YELLOW or RED in the software console. 

Figure 2 - A sample RFID/GPS wrist tag that has a distress button to raise alerts.

The wrist tag also has a push button that can raise an alarm. When the software or the wrist tag alerts are received, an unmanned underwater vehicle is deployed to that position for rescue.

 

Unmanned Underwater Vehicle (Rescue Bot)

An unmanned underwater vehicle is used for rescuing drowning victims. Based on data received from the software console, or if triggered manually by a lifeguard, the UUV sets course to the exact position of the victim based on GPS co-ordinates.

Figure 3 - Illustration of a sample unmanned underwater vehicle with easy grip handles.

The UUV (rescue bot) has easy grip handles for drowning victims to grab, and secure themselves. The vehicle is fitted with search lights and cameras, so that the rescue can be monitored from the console.

Alternately, a lifeguard can disable the automatic mode (UUV-mode) and remotely operate (ROV-mode) the rescue bot.

Note: Detailed specifications and information about the unmanned underwater vehicle is available here.