War Thunder Target Finder (WTTF)

Overview

War Thunder Target Finder (WTTF) is a real-time object detection assistant designed for the military game War Thunder. In game modes like "Realistic Battles" and "Simulator Battles," enemy vehicles are not visually marked, making it challenging for players to spot threats in the environment. WTTF aims to support players by providing a transparent overlay that highlights detected vehicles using a trained YOLO object detection model. The tool enhances situational awareness while preserving the game's immersive experience.

---

Technical Architecture

WTTF is structured into modular components that work together seamlessly:

• User Interface (GUI):
  Provides a simple and user-friendly interface for configuring detection settings such as FPS, bounding box color, and window title.
  Built using tkinter and ttk.

• Detection Control:
  Manages the detection loop in a separate thread, coordinating between GUI, window finder, screenshot capture, detection engine, and overlay.
  Uses Python's built-in threading.

• Window Finder:
  Locates the War Thunder game window by searching for a specified title substring.
  Uses win32gui library.

• Screenshot Capture:
  Captures the visual content of the target window for analysis.
  Uses PIL.ImageGrab.

• YOLO Model (Detection Engine):
  Runs the object detection model (trained YOLOv11) on the captured frames to detect vehicles.
  Built using ultralytics and torch.

• Transparent Overlay UI:
  Creates a transparent, click-through overlay that displays detection results over the game window.
  Uses tkinter.Canvas.

• Utilities:
  Ensures DPI awareness and handles user input errors across the application.

• Data Transformation & Training Scripts:
  Tools for converting detailed vehicle datasets into simplified classes ("vehicle" and "dead") and training the YOLO model.

---

Installation Instructions

1. Go to the release page of the project.
2. Download all the .zip files and extract them to your preferred directory.
3. Download the model file best.pt.
4. Place both the extracted application and best.pt in the same directory.
5. Run the executable file WTTF.exe.

---

Group members and their roles

Xiaoyang Chen (xc52)

Initiating project ideas and assigning weekly tasks

Responsible for screenshot capture, model training and related scripts

Hayes Danaher (hrd2)

Responsible for detection control and utilities

Sonit Maddineni (smadd28)

Responsible for window finder and running YOLO model

Vincent Lavelle (hsule3)

Responsible for user interface and overlay UI

All members

Responsible for collecting and annotating image dataset