Field Surveillance using Unmanned Aerial Vehicle (UAV)

This project aims to assist The Police by helping them detect suspicious activities in agricultural fields where robbers usually hide. A drone will be used to detect suspicious activities in farms and fields where it is usually difficult and tedious to manually do such a thing

Project Title

Field Surveillance using Unmanned Aerial Vehicle (UAV)

Project Area of Specialization

Artificial Intelligence

Project Summary

This project aims to assist The Police by helping them detect suspicious activities in agricultural fields where robbers usually hide. A drone will be used to detect suspicious activities in farms and fields where it is usually difficult and tedious to manually do such a thing due to the size of the field and the dangers involved during nighttime. A thermal camera will be used for this purpose. The microcontroller used for this project will be trained using AI to detect suspicious activities. Once such activity is detected, an alert will be sent to the user along with the captured image. The user will then decide whether it is a false alarm or not and takes action accordingly.

Project Objectives

We intend to develop an Autonomous Quadcopter equipped with the ultimate power dormant by GPS in the field of navigation. GPS provides reliable location and time information. UAVs armed with GPS offer enhanced control in the air with superior observation, surveillance, monitoring & remote sensing abilities. In this project, a Quadrotor is mounted with Pixhawk (flight controller) to control and automate the UAV. Using a companion computer raspberry pi 4/ Nvidia jetson Nano we can transmit the messages via MaVproxy and MAVLink protocols. A thermal camera connected with raspberry pi 4/ Nvidia jetson Nano can stream a live video. Pixhawk is implemented with a sophisticated control algorithm that acts as the brain of the quadrotor with sensory feedback from gyroscope and accelerometer sensors to acquire the desired stability and proper flight. The Quadrotor can fly fully autonomous and cover the specific field area assigned by the user. The CNN (Convolution Neural Network) model with hybrid neural network Long and Short Term Memory (LSTM) can predict the human suspicious activity using the key points within the field and after detection effectively inform the ground station using Telemetry.

Project Implementation Method

First, we built a UAV model with the right electronic components that should satisfy the thrust-to-weight ratio. After the UAV completion, we test the flight and auto-tune it. We established the wireless communication between the flight controller and the ground control station using MAAVlink and MAVProxy protocols using a companion computer Nvidia Jetson Nano. 

Data sets (images) will be collected with different human postures and then we will train the data using a hybrid neural network LSTM. The activity then is detected with the help of keypoints extracted from the image running in the feed-forward loop.  

Benefits of the Project

1.?????? Recent Surge in Crime

An abrupt hype is being observed in the crimes during night time and it is because criminals can easily escape from the spot and hide in such places that are difficult to find and search. The view from a drone can not only provide the location of the criminal, but also an understanding of the surrounding area, offering valuable information such as where the criminal might be headed towards.

2. Conventional methods have become Outdated

The conventional method uses a lot of time and resources and usually fails to catch criminals. The police need a workforce for the search operation which is time-consuming and not effective. Drones allow police to see places and things that fixed cameras cannot. The camera will target and zoomed the suspect to analyze the details more clearly at locations where humans cannot reach. Additionally, there is a myriad of places, like the middle of a field, where there is no vantage point to conduct standard surveillance without being seen. In these cases, a high-flying drone offers the best opportunity to get a solid view of the scene.???????

3. Reduced Risk

During the search operations, especially in areas like the agriculture field suspects can easily hide and can make assault the police official very easily. By using drones we can find the actual location of the suspects and avoid surprise attacks from them hence reducing the chances of injuries and death.???????

4. Easily Deployable

UAVs are also great tools for active criminal situations because they can be deployed from almost anywhere and stored in the trunk of a cruiser. They can also access areas traditional helicopters cannot. UAVs can fly lower to the ground, get into tight spots, hover under bridges and structures, and even fly inside buildings to help the experts gather as much detail as they can.

Technical Details of Final Deliverable

The final prototype will be an autonomous UAV mounted with a camera and companion computer Jetson Nano. The auto mission will be assigned to a UAV to scan the field at an altitude minimum of 12 meters. All aspects of the mission will be observed on Mission Planner. The information of detected activity and location will be observed. Live video streaming of detected activity will be displayed on the web browser which can hold multiple clients as well.

Final Deliverable of the Project

HW/SW integrated system

Core Industry

Security

Other Industries

Agriculture

Core Technology

Artificial Intelligence(AI)

Other Technologies

Internet of Things (IoT)

Sustainable Development Goals

Industry, Innovation and Infrastructure

Required Resources

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