一种数据流驱动的Python视觉实验程序开发方法

A Data Flow-driven Python Visual Experiment Program

摘要:

Python+OpenCV可以满足各种常见的机器视觉实验程序的设计需求,但与MATLAB等商业软件相比,缺乏一个好用的可视化编程环境。为此,提出一种数据流驱动的视觉实验程序开发架构；将按钮、文本框等GUI交互控件和图像显示、图像处理功能封装为带输入、输出的双端口组件；通过组件间输入、输出端口的直连就可以快速形成可运行的视觉实验程序框架；将双端口组件和组件之间的连接导出就可得到框架程序的Python脚本,供用户进一步完善和修改,其功能就类似与MATLAB中的App Designer；为便于理解,最后给出了一个手势检测视觉实验的框架程序快速生成和修改完善样例。

关键词:Python;数据流;可视化编程;机器视觉;实验开发

Abstract:

Python+OpenCV can meet the design needs of various common machine vision experiment programs, but compared with commercial software such as MATLAB, it lacks a useful visual programming environment. To this end, a data flow-driven visual experiment program development architecture is proposed; GUI interactive controls such as buttons and text boxes, image display, and image processing functions are encapsulated into dual-port components with input and output; through input and output ports between components The direct connection can quickly form a runnable visual experiment program framework; export the dual-port components and the connections between the components to get the Python script of the framework program for users to further improve and modify. Its function is similar to that in MATLAB App Designer. In order to facilitate understanding, a framework program for rapid generation and modification of gesture detection visual experiments is finally given.

Key words:Python; Data Flow; Visual Programming; Machine Vision; Experimental Development