Archivedc++YOLOQT
Qt YOLO Inferencer
A desktop inference application for YOLO series models developed with Qt and C++, supporting real-time object detection from images, videos and cameras, with model loading, inference, result visualization and parameter configuration.
Project Highlights
- Developed with Qt C++, native cross-platform desktop GUI, compileable and runnable on Windows and Linux.
- Supports three input sources for YOLO object detection: images, local video files and USB cameras.
- Encapsulates complete YOLO inference pipeline including image preprocessing, model inference, bounding box post-processing and rendering.
- Visual parameter panel to dynamically adjust confidence threshold, NMS IoU threshold and inference resolution.
- Multi-thread design: inference tasks separated from UI rendering to prevent UI freeze and ensure smooth real-time preview.
- Compatible with ONNX YOLO models, directly load exported YOLOv5 / YOLOv8 models, no Python runtime required.
- Draw bounding boxes, class labels and confidence scores in real time, and supports saving inference image/video results locally.