Skip to content

About ATSVisionOps

ATSVISIONOPS.png

To keep up with the growing feature requirements and to provide a better user experience, I have completed a full architectural migration from Tkinter to PySide6. The new version brings significant improvements in performance, stability, and user experience.


🌱 My Original Intention: Breaking the "Last Mile" of Deploying Algorithms

At the very beginning of developing ATSVisionOps, we relied on a scattering of ad-hoc scripts to assist our work. Parameters could not be adjusted in time and were not intuitive, which forced everyone to spend large amounts of time on inefficient, repetitive work:

  • Cumbersome environment setup (CUDA version conflicts, dependency hell)
  • Dull repetitive labor (manually annotating thousands upon thousands of images)
  • Unintuitive parameter tuning (code parameters like a blind box, with no interface)

I realized that this should not be "hard labor". So I decided to build a visual algorithm assistant that is genuinely "ready to use out of the box" and "elegant", so that everyone can bring their focus back to more productive work.


🚀 Development Journey: The Evolution from a Tool into an Ecosystem

(2024/06/29 ~ present)

  • 🟩 Demo v1.0.0 ~ Demo v2.2.3 Exploration phase: The project was initiated; the most basic cloud compute API calls were implemented; the project was still at the exploratory stage.
  • 🟨 Demo v2.3.0 ~ Demo v4.1.0 Gaining direction: The basic direction was established, aimed at solving high-frequency, repetitive, inefficient work.
  • 🟧 Alpha v1.0.0 ~ Beta v1.8.4.5 First refactoring: A UI based on Tkinter was implemented, integrating the features; the ecosystem began to take shape and started to be promoted.
  • 🟥 Stable v2.0.0 ~ Stable v2.5.3 Official release: Optimization and updates to the official Tkinter-based release; the project entered a period of stable promotion.
  • 🟦 Stable v3.0.0 ~ present Second refactoring: A completely new version based on PySide6, achieving major breakthroughs and offering a complete, mature ecosystem.

🔍 Core Value

ATSVisionOps is not just a piece of software; it is our understanding of "efficient production":

  • Standardization: Unified data formats, bidding farewell to scattered folders.
  • Visualization: A fully transparent process, with metric changes grasped in real time.
  • Efficiency: Automation across the entire flow, from data collection to model export.

💡 Thanks to every teammate who has been part of this journey. It is your feedback on Gitee, submitted again and again, that has made this software better!


📋 Version Status Overview

Version Branch Maintenance Status Recommendation Level Notes
PySide6 version 🟢 Actively maintained ⭐⭐⭐⭐⭐ The main version; updated continuously; recommended for all users
Tkinter version 🔴 Maintenance ended ⚠️ Not recommended Archived and no longer updated; kept for historical reference only

Important notes:

  • ✅ The software has a built-in automatic update check

  • 🔄 All new features and new operators are exclusive to the PySide6 version

  • 📢 Existing users are advised to migrate to the PySide6 version as soon as possible to get the latest support


🚀 Quick Start

🖥️ System Compatibility

Platform Supported Version Test Environment Status
Windows 10 / 11 Fully tested 🟢 Fully compatible (preferred platform)
Linux Ubuntu 24.04+ Primary test platform 🟢 Runs stably; distributed as a standard .deb package

💡 Special notes on running under Linux: The Linux version has already achieved feature parity and runs stably. Under Ubuntu, ATSVisionOps is packaged and distributed as a standard .deb package, available on the Release & Download page and installable with sudo dpkg -i opsstd_*.deb. If you need the software for other Linux distributions, please contact the developer directly to obtain it.

📦 Installation and Deployment (Windows)

  1. Automatic configuration (recommended)

    Refer to Dependency Installation.

  2. Manual dependency configuration

    pip install numpy opencv-python pillow scikit-learn tqdm