FACE DETECTION AND CROWD COUNTING USING DRONE (U.A.V.)

Artificial Intelligence and Machine Learning is the future of modern technology in this era. The purpose of this project is to automate the existing technology with features of face detection and counting the number of people by using a drone. The existing system of security companies is manual and

Project Title

FACE DETECTION AND CROWD COUNTING USING DRONE (U.A.V.)

Project Area of Specialization

Artificial Intelligence

Project Summary

Artificial Intelligence and Machine Learning is the future of modern technology in this era. The purpose of this project is to automate the existing technology with features of face detection and counting the number of people by using a drone. The existing system of security companies is manual and difficult in Pakistan. They use security cameras to detect a person’s face within a limited area. But with this system, they can detect a person’s face with an aerial view which covers a lot of area and distance. The security companies wherever they are they must protect the crowd; they do not know how many people are in the crowd so there is the risk in dealing with the crowd. But with this project, one can do this easily with the help of a drone, and the risk is reduced. This project is to solve the problem of security companies with the features of counting the number of people and face detection

The project implements face detection on the drone to automate the process of manually detecting the person’s face and counting the number of people. The project develops this solution for the security and defense-related departments where security is their priority by using the best latest technology. This drone can be sent to a specific environment eliminating ground constraints, limitations and to detect the person’s face, and count the number of people within a specific range. This minimizes the life-related risk, and it is a fast and reliable method of face detection and counting the number of people in difficult scenarios.

Project Objectives

The objectives of this proposed system are:

• To design a system that is mostly used by security companies.
• To develop a system that does face detection and counts the number of people.
• To improve the existing manual system.
• To reduce the security risks and problems.
• To increase efficiency and reliability.
• To validate the existing system that helps the security companies.   

Project Implementation Method

This product can detect any person’s face. Moreover, the specialty of this product is that it is a generic version of the product that can detect any person’s face and the targeted area to count the number of people in a crowd for which the system is trained. For that purpose, the raspberry pi fisheye camera along with a small computer which is known as raspberry pi 4 is attached to the drone which has got all the data and is responsible for detection. For detection ML algorithm is used along with OpenCV to extract the features of the face and then label the matching percentage after comparing the face with provided trained images/video. All the data is received on the GCS (Ground Control System).

The system is designed and modeled in such a way that it has three main modules that are drone (hardware), face detection using machine learning and counting of the crowd using machine learning. Users are assumed to have basic knowledge about how to operate a drone. There are a few steps to fly it. First, the user must plug the battery wire with the drone’s mainboard wire to power up the drone then using the joystick users have to switch the flight mode to loiter or altitude hold mode to arm the drone. Users must move the throttle stick fully down and then move it to the right corner. If land mode is selected or any other pre-arming check is failed, then flight control is not initialized, and the drone does not get armed. Users should know all controls of the transmitter (controller/joystick) to control the drone. User should also have to understand the GCS interface and understands all the controls. The user’s lack of knowledge in any of the stated precautions leads to major failure or damage to the product. The proposed system is run using python and Open-CV on Linux (raspberry pi OS) and Windows OS. Sometimes there are latency and frame drop issues when the connection is weak between GCS and Raspberry pi so the user must have a static and strong internet connection. GCS and vehicle should be on the same network. The proposed system performs face detection by using a face detection algorithm and for other purposes, systems can also perform counting of the crowd by using the specified algorithm. It is assumed that the user should be familiar with the drone flying techniques and should be familiar with the CLI commands and interface.

Benefits of the Project

The purpose of this system is to provide convenience and ease to the security organizations to facilitate them with modern face detection products. This product brings advancement in the security companies with the ease of face detection and counts the number of people which is generally done manually and because it is placed on the ground and cover the very less and specified area but by using drone it can cover a lot of area by hovering over the ground. It reduces security risks and manual efforts for counting the number of people.

Technical Details of Final Deliverable

The technical details of this project are:

• Drone frame F550 for Hexa-Copter.
• ArduCopter APM 2.8 Flight Controller is the brain of the drone. It's a circuit board with a range of sensors that detect the movement of the drone, as well as user commands. Using this data, it then controls the speed of the motors to make the drone/craft move as instructed.
• 10x45 Propellers will be used, which is attached with motors.
• 1400KV brushless outrunner motors are what propellers are connected to which causes them to spin around and generate thrust to enable your drone to fly.
• 30A Electronic Speed Controller (ESC) an electronic circuit that controls a motor’s speed and direction as instructed by Flight Controller.
• 5200mah battery is used which power up the drone.
• 10x45 Propellers are used to hover the drone which is attached with motors.
• M8N Ublox GPS module with a compass is used to maintain and update the location of the drone.
• FlySky FS-T6 digital transmitter and receiver are used to operate the drone. The transmitter transmits the signal to the receiver which is attached to the flight controller.
• Flight controller power module is used, which calculates battery timing and give instructions to flight controller to maintain a flight of drone within safe battery time of the drone.
• HC-SR04+ ultrasonic module is used to avoid collision of the drone with any obstacle.
• Raspberry Pi is a small single-board computer attached to a drone and it is powered by a battery. It contains algorithms for both face detection/recognition and crowd counting.
• Raspberry Pi Camera uses live video streaming for face detection/recognition and crowd counting. 

Final Deliverable of the Project

HW/SW integrated system

Core Industry

IT

Other Industries

Security

Core Technology

Artificial Intelligence(AI)

Other Technologies

Robotics

Sustainable Development Goals

Industry, Innovation and Infrastructure

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com