目标检测演示应用介绍 Object Detection Demo Overview

欢迎使用本演示系统,可快速体验图片、视频和摄像头实时目标检测能力。

This demo helps you quickly experience image, video, and real-time camera detection.

1. 总体介绍 General Overview

本应用基于 Django + YOLOv6 构建,定位为“可视化目标检测演示平台”。

支持三类场景:图片检测、视频检测、摄像头实时检测;并提供检测结果展示与会话内历史记录。

Built with Django + YOLOv6 as a visual demo platform, supporting image, video, and camera detection with history records.

2. 用法 How To Use

Use the top navigation to open Image/Video/Camera pages, then upload files or view live stream.

3. 技术路线 Technical Route

Frontend uses Bootstrap templates; backend uses Django; inference is powered by YOLOv6 with OpenCV and ffmpeg pipeline.

4. 构建过程 Build Pipeline

  1. 准备基础预训练权重(pretrained weights)作为初始化模型。
  2. 以 MS-COCO 数据集进行基础训练或迁移学习,获取通用目标检测能力。
  3. 结合自构建训练数据集(自采集/自标注)继续微调,提升业务场景表现。
  4. 导出最终权重并集成到 Django 演示应用中,完成端到端演示部署。

Pipeline: pretrained initialization -> MS-COCO training -> custom dataset fine-tuning -> deploy final weights into this Django demo.