People identification system with unmanned aerial vehicles
Tutor / Supervisor
Student
Fontarnau Galea, Gerard
Document type
Bachelor thesis
Date
2023
rights
Open Access
Publisher
Universitat Politècnica de Catalunya
UPCommons
Abstract
In this Bachelor's Degree Final Project, a mobile application for person identification using a DJI drone and the DJI Mobile SDK and DJI UX SDK software libraries is designed, implemented, and tested. The application tries to identify a specific person among those that are detected in the image. Our proposal, for this, is that the person who wants to be identified wears a GPS device to merge the information from the people detector with the location information provided by the GPS. The operation of the application mainly involves monitoring an MQTT server where different devices will upload their position using geodetic coordinates. Then, based on these coordinates, the application will perform a coordinate system transformation to obtain the pixel coordinates where the device is located. With these coordinates, the device's position can be overlaid on the video. Additionally, the application includes other visual functionalities such as a manager for photos and videos taken by the drone, a mini-map to visualize the drone's and surrounding devices' positions, a coordinate converter from screen points to geodetic coordinates, and the ability to write device data and drone position and attitude data to a text file. For the development of the application, various Software Development Kits (SDKs) are used, which provide the necessary resources for application development. Some of the SDKs used include the Android SDK, DJI UX SDK, and DJI Mobile SDK, with the latter two being from DJI. The MQTT protocol is used for message exchange between the drone and the different devices in the field. This protocol is based on centralized data exchange on a server and utilizes a publish-subscribe system. Furthermore, three different devices are used to obtain the user's location through GPS and send it to the MQTT server for the drone to access. One of these devices is created using a Raspberry Pi, another is an Android app, and the last one is based on a board from ArduSimple. Upon completing the application development, the project's initial objectives have been successfully achieved. A functional application has been programmed, and a device compatible with the specified requirements has been developed.
