基于人工智能的飞行试验监控画面前面板生成算法设计

Design of Front Panel Generation Algorithm for Flight Test Monitoring Screen Based on Artificial Intelligence

摘要:

飞行试验监控画面前面板是遥测数据可视化的核心载体,针对当前采用“示意图绘制-人工解析-手动编码”的传统开发模式,开发周期长、重复性工作多的问题,设计了一种基于人工智能的监控画面前面板生成算法,首先结合OpenCV轮廓提取与OCR技术,自动解析监控画面示意图中的控件类型、位置及参数含义,生成结构化信息；再通过适配的提示词设计,引导大语言模型生成Python脚本,用于生成无需修改的UI界面程序；最后构建原型系统并选取代表性画面验证,结果表明,该算法能精准提取示意图信息,生成界面无需修改即可使用,单画面开发周期从半天缩短至分钟级,有效提升了监控画面开发效率。

Abstract:

The front panel of the flight test monitoring screen is the core carrier for the visualization of telemetry data. In response to the current traditional development mode of "diagram drawing - manual analysis - manual coding", which has long development cycles and a lot of repetitive work, an AI-based algorithm for generating the front panel of the monitoring screen is designed. Firstly, by combining OpenCV contour extraction and OCR technology, the types, positions, and parameter meanings of the controls in the monitoring screen diagram are automatically analyzed to generate structured information. Then, through the design of appropriate prompts, a large language model is guided to generate Python scripts for generating UI interface programs that do not require modification. Finally, a prototype system is built and representative screens are selected for verification. The results show that the algorithm can accurately extract the information from the diagram, and the generated interface can be used without modification. The development cycle for a single screen is shortened from half a day to minutes, effectively improving the development efficiency of the monitoring screen.