Overview

The work showcased in this repository is an impressive collection of object detection and tracking projects that offer a plethora of functionalities geared towards real-time applications. Designed with flexibility in mind, it allows users to host their systems on a cloud server or utilize their own IP cameras, featuring seamless integration with asynchronous processing. Leveraging advanced algorithms like Deep SORT and YOLO v4, this repository not only emphasizes accurate object detection but also ensures efficient tracking across various scenarios.

In addition to the core features focusing on detection and tracking, the projects provide extensive capabilities for traffic counting and real-time object management. With detailed records and statistical outputs, it serves as a vital tool for developers and researchers looking to delve into object detection technologies and their practical applications in smart environments.

Features

Asynchronous Processing: Utilize your own IP cameras with ImageZMQ for efficient asynchronous video streaming.
Advanced Tracking Algorithms: Incorporate Deep SORT and YOLO v4 for precise tracking, with support for Tensorflow 2.0.
Traffic Counting Capabilities: Count vehicles accurately using techniques that ensure only valid tracking IDs are registered.
Multiple Camera Support: Easily integrate multiple IP cameras into the system for comprehensive coverage.
Customizable Counts: Configure directional counting based on angles and set intervals for recording object counts.
Detailed Intersection Records: Gather thorough data on intersection details, including time, coordinates, and angle for each counted object.
Cloud Hosting: Seamlessly host your application on a cloud server for enhanced accessibility and scalability.
Adaptable Object Detection: Implement tracking and counting for various objects with optional visibility of detection confidence, catering to different use cases.

Leonlok

Multi Camera Live Object Tracking